JP2012515533A - Methods of single cell gene expression for diagnosis, prognosis, and drug discovery target identification - Google Patents

Abstract

Methods for disease diagnosis and prognosis are provided by analyzing the expression of a set of genes obtained by single cell analysis. Classification allows for optimization of treatment and determination of whether to treat with a particular therapy, and how to optimize dose, choice of treatment, etc. Single cell analysis also provides for the identification and development of therapeutics that target mutations and / or pathways of cells in a disease state.

Description

This application claims the benefit of US Provisional Application No. 61 / 205,485, filed Jan. 20, 2009, which is hereby incorporated by reference.

Government Rights This invention was made with government support through federal grant U54 CA 126524 awarded by the National Cancer Institute (USA). The United States government has certain rights in this invention.

In recent years, analysis of gene expression patterns has provided methods for improving diagnosis and risk stratification of many diseases. For example, an unsupervised analysis of exhaustive gene expression patterns reveals molecularly distinct cancer subtypes that are distinguished by widespread differences in gene expression in diseases that have been considered homogenous based on conventional diagnostic methods. The type has been identified. Such molecular subtypes are often associated with various clinical outcomes. Comprehensive gene expression patterns can also be tested for features that correlate with clinical behavior to create a prognostic signature.

  Cancer, like many diseases, is often a result of a variety of signaling pathways that ultimately result in a seemingly similar pathological phenotype, rather than the result of a single obvious cause. It can be regarded as several diseases caused by various abnormalities. Identification of polynucleotides that are differentially expressed in cancerous cells, precancerous cells, or cells that are less likely to metastasize compared to normal cells of the same tissue type provides the basis for diagnostic tools Promoting drug discovery by providing targets for candidate agents, and can further help identify therapeutic targets for cancer therapy that are more compatible with the type of cancer being treated.

  Identification of differentially expressed gene products also helps to understand the progression and nature of complex diseases, for example identifying genetic factors responsible for phenotypes associated with the development of metastatic or inflammatory phenotypes The key to doing. Identification of gene products that are differentially expressed at various stages and in various cell types provides an early diagnostic test and can be a therapeutic target. In addition, differentially expressed gene products can serve as the basis for screening assays to identify chemotherapeutic agents that modulate their activity (eg, their expression, biological activity, etc.).

  The most important thing to stop disease progression and reduce morbidity is early diagnosis of the disease. Analysis of patient samples to identify gene expression patterns provides a basis for the treatment of more specific and rational diseases that can result in reduced side effects compared to conventional treatments. Still further, unnecessary treatment can be avoided by confirming that the lesion is of low risk to the patient (eg, the tumor is benign). In short, the identification of gene expression patterns in disease-related cells can provide the basis for therapeutics, diagnostics, prognostics, therametrics and the like.

  As another example, infectious diseases damage tissues and organs, which cause morbidity and mortality of certain organisms. In the case of influenza A infection, the most common cause of hospitalization and death is infection of lung tissue. However, the very cells that infect influenza and the cells that repair damaged lungs are not understood at the single cell level. Such knowledge can help to identify therapeutic targets for intervention, such as new drugs to prevent viral infections, and new treatments to improve morbidity.

  Many tumors contain a mixed population of cancer cells that can differ with respect to the signaling pathways used for their growth and survival. These cancer cells differ with respect to their response to a particular treatment, and the resistance of a particular population of cancer stem cells contributes to the recurrence after cytotoxic radiotherapy and chemotherapy. Thus, hospital treatment failures may be due in part to resistance to treatment of a particular population of cancer cells.

  The initial shrinkage of the tumor immediately after treatment, often observed, can only reflect the relative sensitivity of one subpopulation of cancer cells that may contain a tumor mass, and in long-term survival May not be important. Thus, the most important clinical variable for assessment of response to treatment and prognosis can be the absolute number of specific populations of cancer cells remaining after treatment, rather than absolute tumor size. Once the differences in signaling pathways used by these different populations of cancer cells within a tumor can be identified, therapeutics can be designed that target individual populations of cells. By targeting the entire population, tumors can be eliminated by treatment with drugs that affect different individual populations.

  As another example, inflammatory bowel disease causes destruction of the normal structure of the intestine, causing problems such as diarrhea, bleeding, and malabsorption. These problems are caused by the disruption of the normal mucosal lining of the stomach. The mucosal lining of the colon consists of crypts, where goblet cells, stem cells, and progenitor cells reside at the base of the crypts, while mature cells, including enterocytes and goblet cells, At the top. For inflammatory bowel disease, it is not clear which cell population is damaged and the signaling pathways required to repair the damaged mucosa.

  Many methods that use small numbers of cells to accurately determine the number and phenotype of cells within a disease lesion include inflammatory bowel disease, infections, cancer, autoimmune diseases such as rheumatoid arthritis, and infections It is a major goal for the prognosis, diagnosis and identification of signal transduction pathways that can be targeted by specific therapies for these diseases. The present invention addresses this problem.

  Compositions and methods are provided for use in single cell gene expression profiling and / or transcriptome analysis. One method provided herein is a method of identifying various cell populations in a heterogeneous solid tumor sample, which randomly distributes individual cells from the tumor to distinct locations. Performing a transcriptome analysis on a plurality of genes of individually distributed cells in the distinct locations; and performing a cluster analysis to identify one or more different cell populations. . In some instances, individual cells are not enriched prior to distribution. Transcriptome analysis can be performed on at least 1000 individual cells simultaneously. Transcriptome analysis can be performed using nucleic acid analysis. Separate locations can exist on a planar substrate. In some embodiments, random dispensing is performed in a microfluidic system. Transcriptome analysis can include analyzing expressed RNA, unexpressed RNA, or both. The transcriptome analysis can be a full transcriptome analysis. Transcriptome analysis can include amplifying RNA using a set of primer pairs. In some embodiments, the primer pair is not a nested primer. Transcriptome analysis can be performed on all or a subset of individual cells simultaneously or substantially in real time. The one or more cell populations can be normal stem cells, normal progenitor cells, normal mature cells, inflammatory cells, cancer cells, cancer stem cells, or non-tumorigenic stem cells.

  Further provided herein are methods for analyzing heterogeneous tumor biopsy material from a subject. The method comprises the steps of randomly distributing cells from the biopsy material to discrete locations; performing a transcriptome analysis on at least 50 genes of the individually distributed cells; and transcriptome data To identify one or more characteristics of the tumor. The steps to be performed can be performed without prior enrichment of cell types. The property to identify can be the presence or number of cancer cells, or the number. The identified property may also be the presence or number of stem cells, early progenitor cells, early differentiated progenitor cells, late differentiated progenitor cells, or mature cells. The identifying property may also be the effectiveness of the therapeutic agent in eliminating one or more of the cells. The identified property may also be the activity of a signal transduction pathway, such as a pathway specific for cancer stem cells, differentiated cancer cells, mature cancer cells, or a combination thereof. The methods disclosed herein can further comprise using the characteristics to diagnose a subject having cancer or a stage of cancer.

  Another method disclosed herein is a method for identifying signal transduction pathways utilized by diseased cells. The method includes randomly distributing cells from a heterogeneous sample; performing a transcriptome analysis on the distributed cells; identifying at least one disease state cell using the transcriptome analysis Performing transcriptome analysis of cells in at least one disease state; (a) cells that are not disease states; (b) cells in different disease states; and (c) transcriptomes of stem cells in disease states A step of comparing; and (i) a cell in a disease state, (ii) a stem cell in a disease state, and (iii) a cell that is expressed in a cell in a different disease state depending on the situation but is not in a disease state Identifying a signal transduction pathway that is not expressed therein, thereby identifying a signal transduction pathway utilized by cells in the disease state. The disease state can be cancer, ulcerative colitis, or inflammatory bowel disease. In some embodiments, signal transduction pathways are necessary for survival of cells in the disease state.

  The present disclosure also provides a method for diagnosing a subject having symptoms. The method comprises randomly distributing cells from a heterogeneous sample; performing a first transcriptome analysis on the distributed cells; using the transcriptome analysis to determine at least one disease state. Comparing the first transcriptome analysis of the cells to the second transcriptome analysis of cells not in the disease state, thereby identifying the cells in the at least one disease state, thereby Diagnosing the presence or absence of symptoms associated with cells in the disease state. The disease state can be breast cancer, colon cancer, ulcerative colitis, or inflammatory bowel disease. Transcriptome analysis can include analyzing expressed RNA, non-expressed RNA, or both. The transcriptome analysis can be a full transcriptome analysis.

  Yet another method provided herein is a method for screening for therapeutic agents. The method comprises exposing a first subject having cells in a disease state to one or more test agents; obtaining a heterogeneous tumor biopsy material from the subject from a region of interest; Performing a transcriptome analysis on at least one individual cell from said heterogeneous tumor biopsy, wherein said biopsy material comprises cells of one or more disease states; and Transcriptome analysis is performed on a transcriptome derived from either (i) a second subject having no diseased cells; or (ii) a first subject prior to the exposing step. And identifying an agent that affects the transcriptome of the cell from a test area that is more similar to the transcriptome of the second subject or the first subject prior to exposure. including . The condition can be breast cancer, colon cancer, ulcerative colitis, or inflammatory bowel disease. The therapeutic agent can be an antibody or antibody fragment, small molecule, nucleic acid (eg, siRNA), RNA, DNA, RNA-DNA chimera, protein, or peptide.

  The present disclosure also provides a method for determining the potential effect of a therapeutic agent on a disease. The method comprises dividing a first population of cells in a disease state into individual locations, where each location includes individual cells; at least one type derived from at least one individual cell Determining the expression level of the nucleic acid or protein, thereby creating an expression signature of the disease state; exposing the second population of cells in the disease state to the agent; Dividing into individual locations, where each location includes individual cells; determining the expression level of at least one nucleic acid or protein from at least one individual cell from the second population And comparing the expression levels from individual cells from the second population against the expression signature of the disease state, thereby determining the effect of the agent on the disease. The exposure step can be performed in vivo. In some examples, the first population and the second population are isolated from a subject, eg, a human. The disease can be cancer, ulcerative colitis, or inflammatory bowel disease. The nucleic acid or protein can be a marker for cancer cells, a marker for cancer stem cells, or both. The expression level can be an mRNA expression level. In some embodiments, determining the mRNA expression level includes the expression or lack of expression of 10 or more nucleic acids. The expression level may also be a protein expression level. The dividing step can include exposing the population of cells to an antibody that specifically binds to a protein present on an individual cell.

  Further provided herein are methods for determining the likelihood of a subject's response to a therapeutic agent. The method includes dividing a population of cells from a subject into individual locations, where each location includes an individual cell, and at least one of the individual cells is a diseased cell; Determining the expression level of at least one nucleic acid or protein from at least one of the individual cells of the disease state, wherein the nucleic acid or protein is the target of a therapeutic agent; and said at least Determining the likelihood of a reaction by the analyte based on the expression level of one nucleic acid or protein. The expression level can be an mRNA expression level. In some embodiments, determining the mRNA expression level includes the expression or lack of expression of 10 or more nucleic acids. The expression level may also be a protein expression level. The dividing step can include exposing the population of cells to an antibody that specifically binds to a protein present on the individual cells. The therapeutic agent can be an anticancer agent.

Another method detailed herein provides a prognostic or diagnostic method that utilizes gene expression from individual cells. The method comprises the steps of dividing cells from a heterogeneous sample into separately addressable locations; lysing individual cells and dividing the resulting lysate into at least two parts; from individual cells, Amplifying mRNA or cDNA derived from; determining a gene expression profile from one of the portions of the lysate (wherein said gene expression profile provides information of the subpopulation); and of the target subpopulation Performing a transcriptome analysis on at least one of the cells. In some methods, at least 10 ² , or at least 10 ³ individual cells are analyzed. Cells can be sorted for expression of at least one cell surface marker. The cells analyzed by the methods disclosed herein can be stem cells, eg, hematopoietic stem cells. The initial sample may contain less than 10 ⁶ cells or less than 10 ⁵ cells. Cells can be sorted for expression of at least one of CD34 and Thy1. In some embodiments, the expression of genes associated with at least one or at least five hematopoietic stem cells is determined. Transcriptome analysis is total transcriptome analysis.

  Further provided herein are methods for classifying stem cells. The method includes the steps of: (a) obtaining a transcriptome profile of a stem cell from a sample; and (b) comparing the obtained transcriptome profile with a transcriptome profile of a reference stem cell. The transcriptome profile may include a data set obtained from proteins associated with at least about 5 types of stem cells. The stem cells to be analyzed can be cancer stem cells, hematopoietic stem cells, intestinal stem cells, leukemia stem cells, or lung stem cells. The sample to be analyzed can include cells from cancer, eg, breast cancer or colon cancer. Transcriptome profile analysis can also include the following additional steps: extracting mRNA from a sample of stem cells; quantifying the level of one or more mRNA species corresponding to stem cell specific sequences; and Comparing the level of one or more mRNA species to the level of said mRNA species in a reference sample.

  A method for collecting data about a transcriptome is also provided herein. The method includes collecting data about the transcriptome using any of the methods described herein and transmitting the data to a computer. The computer can be connected to a sequencing device. Data corresponding to the transcriptome can be further screened after transmission. For example, the data can be stored on a computer readable medium that can be extracted from the computer. The data can be transmitted from the computer to a remote location, for example via the Internet.

  This patent or application contains at least one drawing executed in color. Copies of this patent or patent application publication with colored drawings will be provided by the authority upon request and payment of the necessary fee.

Single-cell gene expression analysis by real-time PCR of human “colorectal cancer stem cells” (EpCAM ^high ) purified from human colorectal cancer tissue xenografted into NOD / SCID mice (tumor # 4m6). In the first experiment (Panel A), 16 single cells were analyzed for expression of 5 genes and 27 replicas were run for each cell-gene combination. In this experiment, each gene-specific primer set was used with the sole exception that each mRNA preparation from an individual single cell was used in three consecutive rows of the reaction matrix and no primer was added to the first three columns. Are used in 9 consecutive rows. The level of gene expression for each individual cell can be visualized as a 3 × 9 block using a color scale. In the second experiment (Panel B), a similar approach was followed, but here 16 single cells were analyzed for the expression of 16 genes and 9 replicas were run for each cell-gene combination. . In this second case, each mRNA preparation from individual single cells is used in 3 consecutive rows of the reaction matrix, and each gene specific primer set is used in 3 consecutive columns. That is, the gene expression level for each individual cell can be visualized as a 3 × 3 block using a color scale. In either case, the assay shows a high level of reproducibility and consistency within each set of replicas. Single-cell gene expression analysis by real-time PCR of human “colorectal cancer stem cells” (EpCAM ^high / CD166 ⁺ cells, xenograft # 8m3 derived). In this figure, each row identifies a single cell and each column identifies a different gene. The intensity of gene expression is indicated using a color code. Here, the darker the red, the stronger the intensity, and the darker the green, the weaker the intensity. This analysis clearly shows that EpCAM ^high / CD166 ⁺ tumor cells can be further divided into different subsets based on their transcriptional repertoire. Most importantly, it shows a high level of coordinated co-expression of genes encoding terminal differentiation markers of colonic epithelium (eg, cytokeratin 20, CD66a / CEACAM1, carbonic anhydrase II, MUC2, Trefoil family factor 3) A subset of cells does not express a gene encoding a candidate intestinal stem cell marker or gene known to be required for stem cell function (eg, hTERT, LGR5, survivin), or is expressed only at low levels; Or vice versa. 3A-B. a: Purification of MTICs 11'ESA ⁺ H2K- from lung cells of NOD / SCID mice with breast tumors. The upper panel gated the H2K-Dapilviable lineage, and the lower left panel gated ESA ⁺ cells to facilitate the gating of CD2441 'cells into the lower right panel. b: Real-time PCR analysis of mRNA levels of HIF1a, Snail2, Zeb2, E-cadherin, vimentin, VEC, CCR7, Lox, Cox2 in MTIC and non-TIC. CT value of real-time PCR analysis for microRNA (miR) levels comparing primary TIC and MTIC. Figures 5A-5D. CD66a as a non-tumorigenic cancer cell marker for breast cancer. Copy number polymorphism analysis of 18 cell samples for thousands of CNVs. Some may be related to genomic instability, causing changes in the properties of pluripotent stem cells. Single cell analysis device, principle. Gene set enrichment analysis of stem cell-related gene expression. Normal HSCs that self-regenerate, leukemia stem cells derived from granulocyte / macrophage progenitor cells (GMP) but not expressed by normal GMPs that do not self-renew, breast cancer stem cells (CSCs) and their non-tumor formation Analyzed in sex offspring (NTG). As shown, these genes accounted for a significantly larger proportion in the CSC gene expression signature. A heat map of overexpressed genes is shown. Schematic of “in silico” sorting for isolation of rare cell subpopulations. Cell populations such as hematopoietic stem cells are sorted into 96-well plates by FACS to contain single cells. Lyse the cells and divide the lysate into two parts. One part of the lysate can be analyzed for the expression of a set of genes and cells can be characterized based on transcription rather than expression of surface proteins. Using this information, selected lysates and / or pooled lysates from similar cells are subjected to total transcriptome analysis. Graphical description of computer data collection, storage, and transfer.

DETAILED DESCRIPTION The method of the present invention provides for the identification of primary cells for characterization of cell populations for disease diagnosis, sensitivity to specific therapeutic interventions, prognostic applications, and identification of new drug targets. Utilize single cell gene expression profiling. Divide a heterogeneous cell sample into spatially separated single cells, and select them according to the situation for the property of interest (which may include surface markers), then lyse and amplify the contents And individually analyze the expression of the gene of interest. Cells analyzed in this way are classified according to the individual cell's genetic signature. Such classification allows an accurate assessment of the cell composition of the test sample.

  Conventional methods for analyzing single cells for diagnostic purposes include the use of Coulter counters and flow cytometry to count the number of cells of a given type. However, these measurements are usually based on the use of antibodies against surface markers and do not assay gene expression or protein expression at the mRNA level. Although there are precedents of single-cell PCR analysis, these are too few to provide useful diagnostic information or to provide the ability to identify subtle or related subpopulations of cells in a tissue Of cells and / or genes. The tissue staining methods used by pathologists have similar drawbacks and rely heavily on qualitative judgment by pathologists. Furthermore, these measurements are limited to measuring a small number of parameters. However, the methods of the present invention can measure at least 10, at least 15, at least 20, at least 50, at least 100, at least 200, at least 300, at least 400, at least 500, or more different parameters. Here, the parameters include mRNA expression, gene expression, protein expression, and may further include cell surface markers combined with mRNA, gene, and / or protein expression.

  Before further describing the present invention, variations of the specific embodiments may be made within the scope of the appended claims and the present invention is not limited to the specific embodiments of the invention described below. It is understood. It is also understood that the terminology used is for the purpose of describing particular embodiments and is not intended to be limiting. In this specification and the appended claims, the singular forms “a”, “an”, and “the” include plural referents unless the context clearly indicates otherwise.

  Where a range of values is indicated, each value between the upper and lower limits of the range (up to one-tenth of the lower limit unit unless the context clearly indicates otherwise) and its stated range It is understood that any other stated or intermediate value within is included within the scope of the invention. The upper and lower limits of these narrower ranges may be independently included in the narrower range, subject to any specific exclusion limit within the specified range, and these are also included in the scope of the present invention. . Where the specified range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the invention.

  Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although any methods, devices, and materials similar or equivalent to those described herein can also be used in the practice or testing of the present invention, the methods, devices, and materials described herein are It describes.

  All publications mentioned in this specification are herein incorporated by reference to describe and disclose the subject matter of the invention described in the publications. The above content can be used in combination with the invention described herein.

Identification and Classification of Cells into Populations and Subpopulations This disclosure is for identifying cell populations and subpopulations and for diagnosing, prognosing, and / or identifying therapeutic targets such as diseases To classification methods using populations and / or subpopulations. Diseases include any type of cancer (including but not limited to solid tumors, breast cancer, colon cancer, lung cancer, leukemia), inflammatory bowel disease, ulcerative colitis, autoimmune disease, inflammatory disease, and Infectious diseases are included. The present disclosure also includes antibodies and nucleic acid probes, or reagents that modulate the biological markers herein, such as antibodies and nucleic acid probes useful for the detection of any of the biological markers described herein. Also provided are reagents and kits for use in performing the methods. The method can also determine the appropriate treatment level for a particular cancer.

Single-cell isolation Single-cell gene expression profiling provides applications in disease diagnosis or prognosis, as well as research tools to identify new drug discovery targets. Control diseases include, but are not limited to, immune-mediated dysfunction, cancer, and the like. In the methods of the present invention, a heterogeneous cell mixture (eg, a tumor needle biopsy, an inflammatory lesion biopsy, synovial fluid, spinal tap, etc.) is converted into spatially separated single cells, eg, Divide into multi-well plates, microarrays, microfluidic devices, or slides randomly or in a specific order. The cells are then lysed, the contents are amplified, and analyzed individually for expression of the gene of interest. Cells analyzed in this way are classified according to the individual cell's genetic signature. Such a classification allows an accurate assessment of the cell composition of the test sample. For this assessment, for example, in determining the identification and number of cancer stem cells in a tumor, the identification and number of cells related to immunity, such as the number and specificity of T cells, dendritic cells, B cells, etc. Applications can be found in making decisions.

  In some embodiments, the cell sample to be analyzed is a primary sample, and the primary sample may be newly isolated, frozen, or the like. However, the cells to be analyzed may be cultured cells. Typically, the sample is a heterogeneous cell mixture that includes a plurality of different cell types, different populations, or different subpopulations (eg, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more cell types, populations, or subpopulations). In some aspects, the sample is a cancer sample from a solid tumor, leukemia, lymphoma, etc., which is a biopsy material (eg, needle biopsy material), a blood sample, etc. for disseminated tumors and leukemia, etc. obtain. The sample may be obtained before diagnosis, or may be obtained during the course of treatment or the like.

  For isolation of cells from tissue, an appropriate solution can be used for dispersion or suspension. Such solutions are generally conveniently supplemented with fetal bovine serum or other naturally occurring factors in combination with low concentrations (typically 5 mM to 25 mM) of acceptable buffers. A balanced salt solution (for example, normal physiological saline, PBS, Hank's balanced salt solution, etc.). Convenient buffers include HEPES, phosphate buffer, lactate buffer, and the like. The separated cells can be collected in any suitable medium that maintains cell viability, usually having a serum cushion at the bottom of the collection tube. Various media are commercially available and can be used depending on the nature of the cells. Examples of the medium include dMEM, HBSS, dPBS, RPMI, Iscove's medium and the like, and are often supplemented with fetal calf serum.

  In some embodiments, the cells in the sample are separated on the microarray. For example, a highly integrated live-cell microarray system can utilize microwells that each fit exactly the size of a single cell (Tokimitsu et al. (2007) Cytometry Part). A 71k 1003: 1010; and Yamamura et al. (2005) Analytical Chemistry 77: 8050, each of which is specifically incorporated herein by reference). Pre-enrichment of the cells of interest, such as by FACS or other sorting, is optional, and in some embodiments, cells from the sample are separated into separate locations without any prior sorting or enrichment. Divide into For example, cells from a sample (eg, blood sample, biopsy material, solid tumor) can be individually isolated at different locations. Typically, for solid tissue samples, the sample is mechanically, chemically, and / or enzymatically separated (eg, by treatment with trypsin or ultrasound). Sample-derived cells can be placed in any cell sorting device (eg, a microfluidic cell sorter) to isolate individual cells, for example, to addressable locations on a planar surface. The planar surface can have nicks, barriers, or other features to ensure isolation of individual cells. Isolated cells can then be analyzed according to the methods herein. It is preferred to divide the cells into different positions. Here, each location contains 1 or 0 cells.

  Cells may be sorted prior to separation, for example by flow cytometry. For example, using FACS sorting or sorting by size difference, the initial concentration of the cell of interest is at least 1,000, 10,000, 100,000, according to one or more markers present on the cell surface, Or even more. Such cells may be sorted according to the presence or absence of cell surface markers that are specific markers of the population or subpopulation of interest.

  When cells are isolated at different locations for analysis, the cells may be sorted using a microfluidic sorter by flow cytometry, microscopy, and the like. The microfabricated fluorescence-activated cell sorter is based on Fu et al. (1999) Nature Biotechnology 17: 1109, and Fu et al. (2002) Anal. Chem. 74: 2451-2457 (each Which is specifically incorporated herein by reference). Samples can be sorted using an integrated microfabricated cell sorter using multilayer soft lithography. This integrated cell sorter includes a variety of microfluidic functions including peristaltic pumps, dampers, switch valves, and input and output wells for sorting cells in a coordinated and automated manner Can be incorporated. The active volume of the actuating valve on this integrated cell sorter can be as small as 1 pL and the light sensing volume can be as small as 100 fL. Compared to conventional FACS instruments, microfluidic FACS offers higher sensitivity, no cross-contamination, and lower cost.

  Individual cells can be isolated at different locations (eg, 96 well plates or microarray addresses) for further analysis and / or manipulation. For example, a cell population containing hematopoietic stem cells (HSC) is selected by FACS analysis using an antibody capable of distinguishing HSC from mature cells. Cells are sorted into 96 well plates, lysed by appropriate methods, and lysates are analyzed by qPCR, microarray analysis, and / or sequencing.

A device for isolation of single cells includes a microfluidic cell sorter. The microfluidic cell sorter isolates living cells from cell debris and sorts the cells from a single cell suspension. Microfluidic devices are used in combination with fluorescent signals from one, two, three, four, five, or more different surface markers (eg, labeled antibodies to target populations or subpopulation markers) These can be done and placed in individual bottles for subsequent genetic studies. Other upstream processes such as digestion of tumors or cell cultures to obtain cell suspensions and staining of cells with fluorescent surface markers can be incorporated into the system. The number of cells analyzed depends on the sample heterogeneity and the expected frequency of the cells of interest in the sample. Typically, at least about 10 ² cells are analyzed and at least about 10 ³ , at least 5 × 10 ³ , at least about 10 ⁴ , at least about 10 ⁵ , at least about 10 ⁶ , at least about 10 ⁷ At least about 10 ⁸ , at least about 10 ⁹ , at least about 10 ¹⁰ , at least about 10 ¹¹ , at least about 10 ¹² , at least about 10 ¹³ , at least about 10 ¹⁴ , at least about 10 ¹⁵ , or Analyze more cells.

  In some examples, the single cell analysis device (SCAD) is modular and can perform the following steps in an integrated and fully automated manner: (1) Tissue digestion. Put the organization in the device's input port. Appropriate enzymes are loaded into the device and flowed to digest the extracellular matrix to obtain a cell suspension. (2) Living from debris, for example, by circulating the digested sample suspension into a microfluidic “metamaterial” (which can divide the fluid flow according to particle size) Cell separation. (3) Staining. The filtered single cell suspension is stained in the compartment of the microfluidic device, if appropriate, using an appropriate surface marker. Staining with up to five different markers can be useful in obtaining a population of highly pure cancer cells. (4) Sorting. The stained single cell suspension is circulated to the next compartment of the microfluidic device to sort the cancer cells from the remaining cells. Various embodiments of the sorter are described in the examples.

Expression Profiling Sorted cells can be individually lysed for analysis of the cell's genome (RNA, DNA) and / or protein composition. For example, mRNA can be captured on a column of oligo-dT beads, reverse transcribed on the beads, processed to detach from the chip, and transferred to a macrowell. DNA or RNA may be preamplified prior to analysis. The pre-amplification can be a pre-amplification of the whole genome or transcriptome, or part thereof (eg, gene / transcript of interest). Polynucleotide samples can be transferred to a chip for analysis (eg, by qRT-PCR) and determination of expression profiles.

  The term “expression profile” is used in a broad sense to include expressed proteins and / or expressed nucleic acids. Nucleic acid samples include a plurality of different nucleic acids or populations of different nucleic acids that may contain expression information about a gene of interest that is determinant of the phenotype of an individual cell. The nucleic acid sample includes RNA or DNA nucleic acid (eg, mRNA, cRNA, cDNA, etc.). An expression profile can be any convenient technique for determining differential gene expression between two samples (eg, quantitative hybridization of mRNA, labeled mRNA, amplified mRNA, cRNA, quantitative PCR, etc. Etc.). A subject or patient sample (eg, a cell or collection thereof, eg, a tissue) is assayed. Samples are collected by any conventional method, as is known in the art. In addition, tumor samples can be collected and tested to determine the relative effects of treatments that produce mortality differences between normal and abnormal cells. A gene / protein of interest is a gene / protein that has been shown to be predictive, including the genes / proteins provided herein. Here, the expression profile may include expression data for 5 types, 10 types, 20 types, 25 types, 50 types, 100 types, or more (including all) of the listed genes / proteins .

  The sample can be prepared by a number of different methods known in the art, for example by isolation of mRNA from a single cell, where the isolated mRNA is known in the field of differential expression. As in the case of amplification or use for the preparation of cDNA, cRNA etc. (eg Marcus, et al., Anal. Chem. (2006); 78 (9): 3084-89 See The sample can be prepared from any tissue (eg, lesion or tumor tissue) taken from the subject. Sample analysis can be used for any purpose (eg, diagnosis, prognosis, classification, treatment tracking and / or development). Cells may be cultured prior to analysis.

  The expression profile may be generated from the initial nucleic acid sample using any conventional protocol. A variety of different ways to create an expression profile are known, such as those utilized in the field of differential gene expression analysis, but one representative and convenient way to create an expression profile The type of protocol is quantitative PCR (QPCR, or QT-PCR). Any methodology available for performing QPCR can be utilized, for example, as described in Valera, et al., J. Neuroncol. (2007) 85 (l): 1-10.

  After obtaining an expression profile from the sample being assayed, the expression profile can be compared to a reference or control profile for diagnosis, prognosis, analysis of drug effects, or other desired analysis. A reference or control profile may be provided or obtained by empirical methods. The resulting expression profile can be compared with one reference / control profile to obtain information about the phenotype of the cell / tissue being assayed. Alternatively, the resulting expression profile can be compared to two or more different reference / control profiles to obtain deeper information about the phenotype of the assayed cell / tissue. For example, the resulting expression profile can be compared to a positive reference profile and a negative reference profile to obtain confirmation information regarding whether the cell / tissue has the phenotype of interest.

  The determination or analysis of the difference value (ie, the difference in expression between the two profiles) can be done using any conventional methodology, where, for example, by comparing digital images of the expression profile Various methodologies are known to those skilled in the array arts, such as by comparing databases of expression data. Patents that describe methods for comparing expression profiles include, but are not limited to, US Pat. Nos. 6,308,170 and 6,228,575, the disclosures of which are incorporated herein by reference. Methods for comparing expression profiles are also described herein.

  A statistical analysis step can then be performed to obtain a weighted contribution of the set of genes. For example, a nearest shrunken centroid analysis is applied as described in Tibshirani et al. (2002) PNAS99: 6567-6572 to calculate the centroid for each class, Thereafter, an average square distance may be calculated between a predetermined expression profile normalized by the intraclass standard deviation and each centroid.

  Classification can be defined probabilistically. Here, the cut-off may be derived empirically. In one embodiment of the present invention, a probability of about 0.4 can be used to distinguish between a sedated patient and an induced patient. More usually, a probability of about 0.5, and a probability of about 0.6 or more can be utilized. A “high” probability may be at least about 0.75, at least about 0.7, at least about 0.6, or at least about 0.5. The “low” probability can be about 0.25 or less, 0.3 or less, or 0.4 or less. In many embodiments, the information obtained as described above for the cell / tissue being assayed is used to predict whether the host, subject, or patient should be treated with the targeted therapy, and It is used to optimize the dose in it.

Characterization of cell populations and subpopulations In some embodiments of the invention, cancers due to expression of cancer stem cell markers (eg, CD66a) when using epithelial cancers including, but not limited to, breast cancer and colon cancer, for example. CSCs can be identified by characterization of stem cells. There are subpopulations of tumorigenic cancer cells that have both self-renewal and differentiation capabilities. These tumorigenic cells are involved in tumor maintenance and give rise to a large number of offspring that are not tumorigenic and exhibit abnormal differentiation, thereby meeting the criteria of cancer stem cells. A subpopulation of cancer cells that differentially express a marker of the invention is capable of tumorigenesis. As shown herein, there is heterogeneity in cell populations that are positive for cancer stem cell marker expression, eg, cells that are negative for CD66 cells here (CD66 ⁻ ) include cancer stem cells (Tumorogenic) is abundant, while CD66a ⁺ cells are not tumorigenic. Detection of such inhomogeneities in the population makes it possible to determine subpopulations.

  One skilled in the art understands that multiple sequences (representing genes, transcripts, and / or proteins) can be analyzed. Such sequences allow for the determination and / or differentiation of the phenotype of the cells in the sample.

  A marker, or marker panel, is based on a number of aspects of a target population or subpopulation in a sample (eg, tissue source (eg, neural or epithelial), or disease state (eg, cancerous or non-cancerous) Other sequences useful for distinguishing cell populations (eg, cancer stem cells derived from normal cells) can be selected for gene alterations (eg, up-regulation or down-regulation) in the target population. It can be determined using the methods described herein, such as by detection.

  Nucleic acids useful in distinguishing one population from another can be up-regulated or down-regulated compared between populations. For example, the expression of some nucleic acids is upregulated or downregulated in cancer relative to normal cells, stem cells relative to differentiated cells, and cancer stem cells relative to differentiated cancer cells. The In some cases, gene up-regulation or down-regulation can be used to distinguish subpopulations within a larger population. For example, some nucleic acids are expressed only in normal cells, expressed in normal cells and cancer cells, or expressed only in cancer stem cells.

  Nucleic acids may be up-regulated or down-regulated compared to another population or sub-population, a specific nucleic acid whose expression level is known, or a standard expression level. Alternatively, when analyzing the expression of a large number of genes, create a heat map by subtracting the average, dividing each gene individually by the standard deviation, and assigning a numerical value based on the degree of deviation from the average can do. For example, a value of +/- 1 may indicate 2.5-3 standard deviations from the mean. Such an analysis can be further refined so that genes in the “+/− 3” range can be used to cluster different types of populations (eg, clustering algorithms (A value of “+3” is given to cancer and a value of “−3” is given to normal tissue so that it can be distinguished). An up-regulated gene may be a “+” value.

  In some examples, a differentially expressed nucleic acid combination can be used as a profile for a particular population or subpopulation. The above profile includes any number of differentially expressed nucleic acids and / or proteins, eg, at least one, two, three, four, five, six, seven, eight, nine , 10 types, 11 types, 12 types, 13 types, 14 types, 15 types, 16 types, 17 types, 18 types, 19 types, 20 types, 25 types, 30 types, 35 types, 40 types, 45 types, 50 One or more types of nucleic acids and / or proteins may be included. In some examples, nucleic acids utilized to identify a target population or subpopulation can be similarly expressed in target and non-target populations, or non-target subpopulations. Such similarly expressed nucleic acids are generally utilized in combination with other differentially expressed nucleic acids to identify the target population or subpopulation.

  The methods described herein can be used to analyze heterogeneous cell populations from any source (eg, biopsy material, normal tissue, solid tumor, etc.). Such methods can be used to determine the presence of a target cell, cancer, or other stem cell, such as any cell population (e.g., a target population, heterogeneous population or subpopulation within a larger heterogeneous population or subpopulation), Alternatively, the entire heterogeneous population can be isolated and used for analysis.

Discovery of Biological Markers The methods disclosed herein can determine new markers associated with a cell population or subpopulation (eg, normal cells, cancer cells, cells in a disease state). Markers can include any biological marker including, but not limited to, DNA, RNA, and protein. In some examples, a cell population marker is a gene or mRNA that is not normally expressed in a given cell (eg, expression of a stem cell gene by a progenitor cell, or a differentiation marker that a cell expresses, or a differentiation marker) Expression of proliferative genes by cells that also express). Typically, two or more markers, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14. 15 types, 16 types, 17 types, 18 types, 19 types, 20 types, 30 types, 40 types, 50 types, 60 types, 70 types, 80 types, 90 types, 100 types, 200 types, 300 types, Evaluate 400, 500, 600, 700, 800, 900, 1000, or more markers. If the marker is expressed RNA, it can be determined from any part of the transcriptome to the entire transcriptome (including the entire transcriptome).

  Analysis of nucleic acid expression patterns in a particular target cell population or subpopulation can lead to the identification of new biological markers that distinguish the target population or subpopulation from others. For example, if a unique surface marker protein is expressed in a target population or subpopulation, isolate and / or identify cells of that population or subpopulation in the same individual or other individuals (eg, by FACS) Antibodies that bind to the marker can be developed for use in the future. Identification of a population or subpopulation specific biological marker includes the absence of a particular marker on the cell population or subpopulation, which allows negative selection. The presence of a marker in a population or subpopulation can be determined using the methods described herein and can be used to define a cell population. The mRNA in the cell population or subpopulation being analyzed indicates that a particular gene is differentially expressed in normal and cancer cells. Differential expression can include an increase or decrease in transcription levels, lack of transcription, and / or changes in expression regulation (eg, different expression patterns in response to stimuli). An mRNA or other marker that is a marker for a cell population or subpopulation can also include mutations present in the cell population or subpopulation (eg, cancer cells and cancer stem cells, but not normal cells). One skilled in the art will appreciate that such a marker may indicate a cell population from one individual tested and / or may indicate a marker for many individuals. In some examples, the expressed mRNA is translated into a protein, which can be detected by any of a variety of protein detection methods (eg, immunoassays, Western blots, etc.).

  Other markers that can be detected include microRNA. In some cases, the expression of certain microRNAs is about 1.5 times, 1.6 times, 1.7 times, 1.8 times, 1.9 times, 2.0 times, 2.1 times, 2.2 times, 2.3 times compared to similar cell populations. MicroRNA expression level is a marker for a cell population when it is increased, decreased, It becomes.

Determination of Transcriptome in Cell Populations and Subpopulations To obtain further information about cells isolated by any of the methods of the invention (eg, FACS separation of cells from the population followed by partial transcription analysis) Further analysis can be useful. In some examples, individual cells isolated from a sample are lysed (eg, by isolation of individual cells with or without prior enrichment) and nucleic acid of interest (eg, genomic DNA, mRNA) Etc.). Utilize gene or gene panel transcriptional analysis to categorize isolated cells into groups that show similarity in their expression profiles (eg, cancer stem cells versus non-stem cells) as described herein it can. While not wishing to be bound by theory, such information suggests a difference in function as a gene and may state that the cell is closely related to its function. Once the cells are organized into groups of similar cells (eg, cells displaying similar or the same transcription profile), contain lysates from individual cells and / or pooled nucleic acids from similar cells The resulting lysate can be further analyzed at the transcriptome level. In some cases, a lysate (eg, a single cell or a pool of similar cells) is used to define a portion of the transcriptome of each cell and / or pool of similar cells (eg, high-throughput sequencing). ). By comparing and / or combining results from individual cells with results from other similar cells, transcriptome information from individual cells can be analyzed at the population level. Transcriptome information from pools of similar cells can also be used to define the transcription characteristics of such pools.

  Any cell population (eg, a cell population comprising stem cells) can be studied in such a manner. In some embodiments, the cells include embryonic stem cells, adult stem cells (including but not limited to cancer stem cells, hematopoietic stem cells (HSC), and mesenchymal stem cells), and stem cells that include induced pluripotent stem cells. Is included. In general, a cell population is a heterogeneous population (eg, a clinical specimen). A subpopulation of interest in a larger cell population can be isolated by any method herein (eg, FACS sorting) according to any relevant criteria (eg, surface protein expression). . In some embodiments, the cells so sorted are each in a sorted population of 10 or fewer cells, 5 or fewer cells, 4 cells, 3 cells, 2 cells. Or into a compartment to contain one cell.

  In some embodiments, the cells are lysed and divided into two or more parts. One portion of the lysate is further analyzed to detect and / or differentiate a subpopulation within a larger heterogeneous population (eg, analysis of a gene panel to detect expression). Cell-derived lysates that have been shown to be in the subpopulation of interest (eg, hematopoietic stem cells) are further analyzed. Lysates from individual cells or pooled lysates from similar cells can be analyzed. The `` similar cell '' population is determined by 1 type, 2 types, 3 types, 4 types, 5 types, 6 types, 7 types, 8 types, 9 types, 10 types, 15 types, 20 types, 25 types, 30 Type, 35 types, 40 types, 45 types, 50 types, 55 types, 60 types, 65 types, 70 types, 75 types, 80 types, 85 types, 90 types, 100 types, 200 types, 300 types, 400 types, This can be done based on the similarity in expression of 500, 600, 700, 800, 900, 1000, or more genes.

  The cells and cell populations or subpopulations of interest can be further analyzed. A cell population or subpopulation includes cells that are part of the original sample (eg, 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8% of the original sample). , 9%, 10%, or more cells). Using the methods described herein, 51%, 52%, 53%, 54%, 55%, 56%, 57%, 58%, 59%, 60% of the isolated population or subpopulation 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%, 74%, 75%, 76%, 77 %, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, The target cell population or subpopulation must be defined so that 94%, 95%, 96%, 97%, 98%, 99%, or 100% do not contain cells that are not members of the target population or subpopulation. It can be isolated from a homogeneous sample. If lysates are prepared from cells isolated from the original population, studies of similar cell populations can be performed by pooling lysates from similar cells.

  Further analysis of cells, populations, and / or subpopulations can include whole transcriptome analysis. In some examples, the lysate contains mRNA that can be amplified for analysis (eg, cDNA) or is analyzed directly (eg, mRNA sequencing, microarray analysis). Amplification of mRNA can be performed by any method known in the art (for example, in vitro transcription, ligation-PCR cDNA amplification). In some embodiments, mRNA amplification can be performed in or using a microfluidic device. Whole transcriptome analysis can be performed by sequencing platforms such as those commercially available from Illumina (RNA-Seq) and Helicos (Digital Gene Expression or “DGE”). In some embodiments, the polynucleotide of interest is sequenced. The target nucleic acid may be sequenced by, for example, a conventional gel electrophoresis based method using Sanger type sequencing. Alternatively, sequencing may be performed through the use of several “next generation” methods. Such “next generation” sequencing methods include, but are not limited to: (1) Margulies et al., Nature (2005) 437: 376-380 (2005); US Patent Nos. 7,244,559; 7,335,762; 7,211,390; 7,244,567; 7,264,929; 7,323,305; including but not limited to the methods and apparatus described in U.S. Pat. (2) U.S. Patent Application No. 11/167046 and U.S. Patent No. 7501245; U.S. Patent No. 7491498; U.S. Patent No. 7,276,720, and U.S. Patent Application Publication No. US20090061439; US 20080087826; US20060286566; US20060024711; US20060024678; US20080213770; and US20080103058, sold by Helicos BioSciences Corporation (Cambridge, MA) (3) Items sold by Applied Biosystems ( (E.g., SOLiD sequencing); (4) those sold by Dover Systems (e.g., Polonator G.007 sequencing); (5) U.S. Patent Nos. 5,750,341; 6,306,597; and 5,969,119. Such as those sold by Illumina; and (6) U.S. Patent Nos. 7,462,452; 7,476,504; 7,405,281; 7,170,050; 7,462,468; 7,476,503; 7,315,019; 7,302,146; 7,313,308; and U.S. Patent Application Publication Nos. US20090029385; US20090068655; US20090024331; and US20080206764. What All references are incorporated herein by reference. Such methods and apparatus are provided herein by way of example and should not be construed as limiting.

  Whole transcriptome analysis can be performed for a number of reasons that allow further characterization of various cell subpopulations including, but not limited to: (1) the unique biological of the subpopulation Detecting the activity of genes that can reveal properties and / or transcription factors that control their development; (2) can be used to purify subpopulations (eg by FACS sorting) Localization and / or characterization of possible surface markers; and (3) distinguishing subpopulations from the entire population (eg, cancer stem cells vs. normal tissues), as potential drug targets for disease, cellular genes And / or detecting and / or characterizing the gene product.

  Analysis of populations and / or subpopulations (eg, by transcriptome analysis) can allow improved techniques for isolating cells belonging to the subpopulations. For example, when subpopulation specific surface antigens are revealed by the above method, antibodies developed by any available antibody synthesis method can be used to remove such cells from a heterogeneous population (eg, patient sample). It can be isolated. In addition, transcriptome profiles can be used to develop gene expression panels that can be used to identify cells from other populations (eg, samples from the same patient or different patients).

Diagnostic Methods and Prognosis The present invention may find use in prevention, treatment, detection, or research for any condition including cancer, inflammatory diseases, autoimmune diseases, and infectious diseases. Examples of cancer include prostate cancer, pancreatic cancer, colon cancer, brain cancer, lung cancer, breast cancer, osteosarcoma, and skin cancer. Examples of inflammatory symptoms include irritable bowel syndrome and ulcerative colitis. Examples of autoimmune diseases include Crohn's disease, lupus erythematosus, and Graves' disease. For example, the present invention finds use in the following prevention, treatment, detection, or research: gastrointestinal cancers such as anal cancer, colon cancer, esophageal cancer, gallbladder cancer, stomach cancer, liver cancer, and rectal cancer; penile cancer, Urogenital cancers such as prostate cancer and testicular cancer; gynecological cancers such as ovarian cancer, cervical cancer, endometrial cancer, uterine cancer, fallopian tube cancer, vaginal cancer, and vulvar cancer; hypopharyngeal cancer, pharyngeal cancer, Oropharyngeal cancer, lip cancer, mouth and oral cancer, salivary gland cancer, gastrointestinal cancer, and sinus cancer head and neck cancer; metastatic cancer; sarcoma; skin cancer; bladder cancer, kidney cancer, And urinary tract cancers, including urethral cancer; endocrine cancers such as thyroid cancer, pituitary cancer, and adrenal cancer, and islet cancer; and childhood cancer.

  Also provided are methods for optimizing treatment by selecting an initial classification of individual cells in a sample and the appropriate treatment, dose, treatment, etc. based on that classification. This optimizes the difference between the delivery of anti-proliferative treatments to undesirable target cells while minimizing undesirable toxicity. The treatment is optimized by selecting a treatment that minimizes undesirable toxicity while providing effective antiproliferative activity. The treatment can be selected to affect only one subset of cells in the sample. In some cases, less than about 5%, less than about 1%, less than about 0.5%, less than about 0.2%, less than about 0.1%, less than about 0.05%, less than about 0.02%, less than about 0.01% of the cells in the sample Choose a treatment that affects, or fewer cells.

  A signature for a symptom can mean an expression pattern of one or more genes or proteins in a single cell that indicates the presence of one symptom. A cancer stem cell signature can refer to an expression pattern of one or more genes and / or proteins whose expression is indicative of the phenotype of the cancer stem cell. An autoimmune cell or inflammatory cell signature refers to a gene and / or protein whose expression is indicative of an autoimmune cell or inflammatory cell signature. Signatures can be obtained from all or part of the data set, and typically include at least about 5 genes and / or proteins, at least about 10 genes and / or proteins, at least about 15 genes. And / or protein, at least about 20 genes and / or proteins, at least about 25 genes and / or proteins, at least about 50 genes and / or proteins, at least about 75 genes and / or proteins, at least About 100 genes and / or proteins, at least about 150 genes and / or proteins, at least about 200 genes and / or proteins, at least about 300 genes and / or proteins, at least about 400 genes Preliminary / or proteins, at least about 500 genes and / or proteins or, according to more genes and / or proteins, including expression information of genes and / or proteins. When using a subset of data, the subset may contain up-regulated genes, may contain down-regulated genes, and may contain combinations of these. is there.

Analysis of patient samples for clinical application The following description focuses on cancer stem cells, but using the methods described herein, normal cells of any tissue (eg, normal stem cells, normal Progenitor cells and normal mature cells), virus-infected cells, inflammatory cells, progenitor cells, cancer cells (eg, tumorigenic cells, non-tumorigenic cells, cancer stem cells, and differentiated cancer cells), disease states Isolated and / or any cell population including, but not limited to, cells (eg, cancer cells, inflammatory bowel disease cells, ulcerative colitis cells, etc.), microbial (bacteria, fungi, protist) cells, etc. Or can be analyzed. Thus, the details provided using cancer stem cells (CSC) are examples of analyzes that can be performed for any disease state or condition.

  In some embodiments of the invention, the number of CSCs in a patient sample can be determined relative to the total number of cancer cells. For example, cells from a biopsy sample are isolated and analyzed for expression of one or more mRNAs and / or proteins that are indicative of cancer cells, and cells that exhibit a CSC phenotype are quantified. Alternatively, data collected for a particular population or subpopulation of CSCs can be used to perform an affinity (eg, antibody) screen for the population or subpopulation, using such an affinity screen. The number of cells can be quantified. Typically, a higher percentage of CSC indicates a continuous self-renewal ability of cells with a cancer phenotype. Quantification of CSCs in patient samples can be compared to positive and / or negative reference samples (eg, patient samples such as blood samples, remission patient samples, etc.). In some embodiments, CSC quantification is performed during the course of treatment, where the number of cancer cells and the percentage of such cells that are CSCs are measured before and during the course of therapy. And quantified as follow-up. Desirably, treatment targeting cancer stem cells results in a reduction in the total number and / or percentage (%) of CSC in the patient sample.

  CSCs can be identified by their phenotype for specific markers and / or by their functional phenotype. In some embodiments, CSCs are identified and / or isolated by combining cells with reagents specific for the marker of interest. The cell to be analyzed may be a living cell, a fixed cell or an embedded cell.

  The presence of CSC in a patient sample can also be an indicator of the stage of cancer (eg, leukemia, breast cancer, prostate cancer). In addition, CSC detection can be used to monitor response to therapy and to aid in prognosis. The presence of CSC can be determined by quantifying cells having a stem cell phenotype. In addition to determining the cell surface phenotype, it may be useful to quantify cells in a sample having “stem cell” properties. This can be determined by functional criteria such as the ability to self-renew in vivo (eg, xenograft model), the ability to generate tumors.

  The clinical sample used in the method of the present invention can be obtained from various sources, particularly blood, but in some cases, samples such as bone marrow fluid, lymph fluid, cerebrospinal fluid, synovial fluid may be used. Samples can include biopsy materials or other clinical specimens containing cells. Some samples include solid tumors or parts thereof. When assaying cell clumps, such clumps can be separated by any suitable technique known in the art (eg, enzymatic digestion, physical separation). Such samples can be separated prior to analysis by centrifugation, suspension separation, density gradient separation, apheresis, affinity selection, panning, FACS, centrifugation using Hypaque, etc. Use fraction (PBMC). In this manner, individual cells from a sample (eg, a solid tumor) can be analyzed for differential gene expression and / or transcriptome analysis, as described herein.

Once a sample is obtained, it can be used directly, can be frozen, and can be maintained in a suitable medium for a short time. Various media can be used to maintain the cells. The sample can be obtained by any convenient technique (eg, biopsy, blood collection, venipuncture, etc.). In some embodiments, the sample comprises at least about 10 ² cells, more usually at least about 10 ³ , 10 ⁴ , 10 ⁵ , or more cells. Typically, the sample is from a human patient, but also finds use in animal models (eg, horses, cows, pigs, dogs, cats, rodents (eg, mice, rats, hamsters), primates, etc.) obtain.

  Appropriate solutions can be used for dispersing or suspending cell samples. Such solutions are generally conveniently supplemented with fetal bovine serum or other naturally occurring factors in combination with low concentrations (typically 5 mM to 25 mM) of acceptable buffers. A balanced salt solution (for example, physiological saline, PBS, Hank's balanced salt solution, etc.). Convenient buffers include HEPES, phosphate buffer, lactate buffer, and the like.

  Analysis of cell staining can be performed using conventional methods. Techniques that provide accurate counting include fluorescence activated cell sorters. It can have varying degrees of sophistication, for example, it can have multiple color channels, low angle light scatter detection channels and obtuse angle light scatter detection channels, impedance channels, and the like. By utilizing dyes that associate with dead cells (eg, propidium iodide), cells can be selected for dead cells.

  The affinity reagent can be a specific receptor or ligand of the cell surface molecule shown above. In addition to antibody reagents, peptide-MHC antigen and T cell receptor pairs (such as peptide ligand and receptor, effector and receptor molecule, etc.) can be used. Antibodies and T cell receptors may be monoclonal or polyclonal, and may be transgenic animals, immunized animals, human or animal immortalized B cells, DNA encoding antibodies or T cell receptors It can be produced by cells transfected with a vector or the like. Details of the preparation of antibodies for use as specific binding members and their eligibility are well known to those of skill in the art.

  One approach is the use of antibodies as affinity reagents. Conveniently, these antibodies can be conjugated with a label used for separation. Labels to allow easy separation of specific cell types include magnetic beads (which allow direct separation), biotin (which can be avidin or streptoid bound to a support). Including any label known in the art including, but not limited to, fluorescent dyes (which can be used with a fluorescence activated cell sorter), and the like. Fluorescent dyes that find use include phycobiliproteins (eg, phycoerythrin and allophycocyanin), fluorescein, and Texas Red. In many cases, each antibody is labeled with a different fluorescent dye to allow independent selection for each marker.

  The antibody can be added to the cell suspension and incubated for sufficient time to bind to available cell surface antigens. This incubation is usually at least about 5 minutes, usually less than about 30 minutes. It is desirable to have a sufficient concentration of antibody in the reaction mixture so that the separation efficiency is not limited by lack of antibody. The appropriate concentration is determined by titration. The medium from which the cells are separated is any medium that maintains cell viability. One medium that can be utilized is phosphate buffered saline containing 0.1% to 0.5% BSA. Various media are commercially available and can be used depending on the nature of the cells. This often includes fetal bovine serum, BSA, HSA, etc., Dulbecco's Modified Eagle Medium (dMEM), Hank's Basic Salt Solution (HESS), Dulbecco's phosphate buffered saline (dPBS), RPMI, Iscove's medium, PBS containing 5 mM EDTA, etc. are included. Labeled cells can then be quantified for expression of cell surface markers as described above.

  Comparison of differential progenitor analysis obtained from patient samples with reference differential progenitor cell analysis can be done by using appropriate inference protocols, AI systems, statistical comparisons, and the like. An indication of the stage of the disease can be obtained by comparison with analysis of reference differential progenitor cells derived from normal cells, cells from similar diseased tissue, and the like. A database of reference differential progenitor cell analysis can be accumulated. One analysis of particular interest tracks patients to observe early stage of disease acceleration, eg, in chronic leukemia or pre-leukemia stage of disease. The method of the present invention provides for the detection of accelerated events prior to the onset of clinical symptoms and thus early therapeutic intervention (eg, initiation of chemotherapy, increased dose of chemotherapy, altered chemotherapeutic drug selection, etc. ).

Tumor classification and patient stratification Methods are also provided for optimizing treatment by initial classification and by selection of appropriate treatments, doses, treatments, etc. based on that information. This optimizes the difference between the delivery of anti-proliferative treatments to undesirable target cells while minimizing undesirable toxicity. The treatment is optimized by the choice of treatment that minimizes unwanted toxicity while providing effective anti-proliferative activity.

  In one embodiment, the present disclosure classifies lesions (eg, tumor foci, immune disease samples, etc.), thereby grouping patients according to single cell (including CSC) gene expression signature, ie “stratified” Provide a method. For example, tumors classified as containing a high percentage of cancer stem cells are at higher risk for metastasis and death, and can be treated more aggressively than more benign types of tumors. Thus, analysis of populations or subpopulations present in patient samples can be utilized to characterize disease states, monitor treatment regimens, and / or develop therapeutic approaches.

  For clinical trials, each patient sample in the pool of potential patients can be classified as described above. Patients with similarly classified lesions can then be selected for participation in therapeutic drug studies or clinical trials where a population of allogeneic patients is desired. Patient classification can also be used in assessing the effects of therapeutic agents in heterogeneous patient populations. Thus, comparison of individual expression profiles to population profiles for disease classification is safe for a particular patient or patient population (ie, a group of patients having the same type of cancer). Yes, allowing selection or design of drugs or other treatment regimens that are expected to be effective. Classification: 1 type, 2 types, 3 types, 4 types, 5 types, 6 types, 7 types, 8 types, 9 types, 10 types, 11 types, 12 types, 13 types, 14 types, 15 types, 16 types For the expression (or absence of expression) of 17 types, 18 types, 19 types, 20 types, 25 types, 30 types, 35 types, 40 types, 45 types, 50 types, or more nucleic acids and / or proteins Can be done on the basis.

Diagnosis, prognosis, therapeutic evaluation (therametrix), and disorder management The classification methods described herein, as well as their gene products and corresponding genes and gene products, include disease pathways (eg, carcinogenic pathways, Genetic or biochemical markers (e.g. in blood or tissue) that may detect initial changes during inflammatory pathways, etc., and / or monitor the effectiveness of various therapeutic and prophylactic interventions ) As a specific target.

  Staging is the process used by physicians to describe how advanced a patient's cancerous condition is. Staging helps doctors determine prognosis, plan treatment, and evaluate the outcome of such treatment. Staging systems vary depending on the type of cancer, but generally include the following “TNM” systems: the type of tumor indicated by T; the cancer indicated by N, in the immediate vicinity of the lymph nodes Whether it has metastasized; and whether the cancer has metastasized to a more distal part of the body, as indicated by M. In general, if cancer is detectable only in the area of the primary lesion and has not spread to any lymph nodes, it is called stage I. If it has spread only to the nearest lymph node, it is called stage II. In stage III, cancer generally spreads to proximal lymph nodes near the site of the primary lesion. Cancer that has spread to distal parts of the body, such as the liver, bones, brain, or other parts, is called stage IV and is the most advanced stage.

  The methods described herein can facilitate fine-tuning of the staging process by identifying the aggressiveness of cancer (eg, the possibility of metastasis), as well as its presence in various regions of the body. it can. Thus, stage II cancers with a classification as indicating a cancer with a high probability of metastasis can be used to change stage II tumors on the border to stage III tumors, which Justify more aggressive treatment. Conversely, the presence of polynucleotides that exhibit lower metastatic potential allows for a more modest staging of tumors.

  For example, breast cancer biopsies from stage II patients are analyzed by the methods described herein. Breast cancer may be classified as having a high metastatic potential depending on the expression profile determined from the patient's cells. Thus, the treating physician can use such information to treat the patient more aggressively than if no further classification was performed. Determination of the expression of specific markers can also provide information about potential targets in drug therapy (eg, tumorigenic cells from patients expressing drug discovery targets).

Therapeutic Development and Identification The methods and compositions described herein can be used to develop or identify new therapeutic agents and / or to improve existing therapies. For example, single cell analysis is used to analyze expression profiles for target cell populations (eg, cancer stem cells, cancer cells differentiated from cancer stem cells, or differentiated cancer cells) to detect potential targets for therapeutic agents be able to. Potential targets include, but are not limited to, specific biological markers and misregulated pathways. Targets of interest can include markers or pathways specific to the cell population of interest.

  In one example, in order to detect new biological markers that can be targeted for treatment, cells of a target population or subpopulation are expressed for expression of nucleic acids as described herein. Can be analyzed. For example, certain cell surface molecules that are expressed only in cancer stem cells and / or differentiated cancer cells can be converted to potential therapeutic agents (eg, antibodies or other binding moieties (toxins or Can be studied as a target)) that may bind to other such effectors. In other examples, target cell populations can be analyzed for misregulation of pathways involved in disease processes (eg, loss of control of cell cycle mechanisms in cancer cells). Pathways can include, but are not limited to, activators and / or repressors of gene expression, expression of specific genes or sets of genes, and more complex exhaustive pathways. A therapeutic agent that targets such misregulation can potentially affect the target cell to alter the expression of the nucleic acid associated with the target cell. Changes in expression induced by a therapeutic agent can result in up- or down-regulation of nucleic acids. In some examples, treatment of a cell and / or subject with one or more therapeutic agents may result in expression of a nucleic acid that mimics expression in a cell that is not a disease state (eg, treatment may result in non-cancerous treatment). Expression of cell cycle-related genes similar to that of sex cells occurs).

  The methods and compositions described herein can be used to analyze a target cell population for changes in the expression of one or more nucleic acids. The development of new and / or improved therapeutics involves analyzing target cell populations (eg, colon cancer stem cells, breast cancer cells, etc.) and analyzing nucleic acids that exhibit altered expression profiles compared to “normal” cells. Determining. Such cells are tested for their effects on the expression of these and / or other nucleic acids by exposing the isolated cells of the target population to candidate agents and testing for changes in expression of the genes after exposure. Can be used to screen for potential therapeutic agents.

  The methods disclosed herein also include gene expression, pathway function (eg, cell cycle, TERT pathway, oxidative stress pathway), and / or specific types, including but not limited to cell types or forms It can also be used to analyze the effects of compounds that affect cell phenotype. In this way, compounds that affect such phenotypic characteristics can be analyzed in addition to or instead of analyzing the ability of the compound as a therapeutic agent. For example, analysis of gene expression changes in target populations exposed to one or more test compounds (eg, normal colon cells, normal breast cells, cancer cells, stem cells, cancer stem cells, etc.), gene expression or other This can be done to analyze the effect of the test compound on the desired phenotype (eg, marker expression, cell viability). Such an analysis may be useful for multiple purposes (eg, cell cycle studies, or analysis of known or unknown pathways).

  The agent that is analyzed for potential therapeutic value can be any compound, small molecule, protein, lipid, carbohydrate, nucleic acid, or other agent that is suitable for therapeutic use. Exposing the isolated cells of the target population to a library of potential therapeutic agents (eg, antibody libraries, small molecule libraries) to determine effects on gene expression and / or cell viability it can. In some cases, the candidate therapeutic specifically targets the cell population of interest. For example, when a single cell analysis reveals the presence of a mutation present in a target cell (eg, cancer stem cell and / or differentiated cancer cell), the candidate therapeutic can be targeted to that mutation. it can. In some examples, treated cells may be exposed to single cell analysis to determine the effect of a candidate therapeutic on the expression of one or more genes of interest and / or the effect on the transcriptome. it can.

  In other embodiments of the invention, the agent is directed against a cell population or subpopulation of the disease state by specifically binding the agent to a single marker or combination of markers present on the target population or subpopulation. Target. In some embodiments, the agent includes an antibody or antigen-binding derivative thereof specific for one type of marker or combination of markers that may be conjugated to a cytotoxic moiety. Such an approach can be used to deplete a target population or subpopulation (eg, deplete a cancer stem cell population) in a patient.

Therapeutic drug screening assays
Cells expressing a single marker or combination of markers (eg, cells in a disease state) are also in vitro assays for detecting factors and chemotherapeutic agents that act on differentiated cancer cells and / or cancer stem cells, and Useful for screening. Of particular interest are screening assays for agents that act on human cells. A wide variety of assays can be used for this purpose, including immunoassays for protein binding, cell proliferation, differentiation, and determination of functional activity; factor production, etc. (e.g. , Balis, (2002) J. Nat'l Cancer Inst. 94: 2; 78). In other embodiments, an isolated polypeptide corresponding to one marker or marker combination of the present invention is useful in drug screening assays.

  In screening assays for biologically active agents, antiproliferative drugs, etc., one marker or target cell composition is contacted with the target agent and the effect of the agent is output to the cell. It is assessed by monitoring parameters (eg, marker expression, cell viability, etc.); or the binding potency or effect of the polypeptide on enzymatic or receptor activity. For example, a breast cancer cell composition known to have an expression profile of “cancer stem cells” is exposed to a test agent, and the exposed cells are individually analyzed as described herein to It is determined whether the agent has altered the expression profile compared to untreated cells. Any isolated cell population described herein or produced by the methods described herein may be newly isolated, cultured, genetically altered, etc. is there. A clonal culture in which the cells are, for example, divided into separate cultures and grown under different conditions, for example with or without drugs, in the presence or absence of cytokines or combinations thereof It may be a mutant induced by the environment. The manner in which cells respond to agents (eg, peptides, siRNA, small molecules, etc.), particularly pharmacological agents, including the timing of the reaction, is an important reflection of the physiological state of the cell.

  A parameter is a quantifiable component of a cell, particularly a component that can be accurately measured, for example, in a high-throughput system. Parameters can be any cellular component or cell generation, including cell surface determinants, receptors, proteins, or their configuration or post-translational modifications, lipids, carbohydrates, organic or inorganic molecules, nucleic acids (eg, mRNA, DNA, etc.) Or a portion derived from such cellular components or a combination thereof. For example, in one embodiment, cells isolated as described herein are contacted with one or more agents to determine the expression level of the nucleic acid of interest. For example, if the cell exhibits an expression pattern more similar to a cell that is not in a disease state, agents that alter the expression of the detected nucleic acid can be further analyzed for therapeutic effects. Most parameters (eg, mRNA or protein expression) provide a quantitative readout, but in some cases semi-quantitative or qualitative results are acceptable. The readout may include one determined value, and may include an average, median, deviation, or the like. It is characteristic that a range of parameter readouts is obtained for each parameter by multiple identical assays. Variations are expected and a range of values for each set of test parameters is obtained using standard statistical methods along with the general statistical methods used to obtain a single value.

  Agents to be screened include known and unknown compounds including numerous chemical classes, major organic molecules that may contain organometallic molecules, gene sequences, and the like. One important aspect of the present invention is to evaluate candidate drugs, including toxicity tests and the like.

  In addition to complex biological agents, candidate agents include organic molecules that contain functional groups necessary for structural interactions, particularly hydrogen bonding, and are typically at least one amine, carbonyl , Hydroxyl or carboxyl groups, often containing at least two functional chemical groups. The candidate agent may comprise a cyclic carbon or heterocyclic structure and / or an aromatic or polycyclic aromatic structure substituted with one or more of the functional groups. Candidate agents can also be found among biological molecules including peptides, polynucleotides, sugars, fatty acids, steroids, purines, pyrimidines, derivatives, structural analogs, or combinations thereof. In some examples, the test compound may have a known function (eg, mitigation of oxidative stress) but acts through an unclear mechanism or against an unclear target There is also the possibility of acting.

  Drugs with pharmacological activity, molecules with genetic activity, etc. are included. Target compounds include chemotherapeutic drugs, hormones, or hormone antagonists. Examples of pharmaceutical agents suitable for the present invention are those described below: "The Pharmacological Basis of Therapeutics", Goodman and Gilman, McGraw-Hill, New York, New York, (1996), 9th edition, following sections: Water, Salts and Ions; Drugs Affecting Renal Function and Electrolyte Metabolism; Drugs Affecting Gastrointestinal Function; Chemotherapy of Microbial Diseases; Chemotherapy of Neoplastic Diseases; Drugs Acting on Blood-Forming organs; Hormones and Hormone Antagonists; Vitamins, Dermatology; and Toxicology (all incorporated herein by reference). Toxins, as well as biological and chemical warfare agents are also included. See, for example, Somani, S.M. (eds.), “Chemical Warfare Agents”, Academic Press, New York, 1992.

  Test compounds include all of the above classes of molecules, and may also include samples whose content is not clear. Of interest are complex mixtures of naturally occurring compounds from natural sources such as plants, fungi, bacteria, protists, or animals. Many samples will contain the compound in solution, but solid samples that can be dissolved in a suitable solvent can also be assayed. Samples of interest include environmental samples (eg, groundwater, seawater, mining waste, etc.), biological samples (eg, lysates prepared from crops, tissue samples, etc.); manufacturing samples (eg, As well as libraries of compounds prepared for analysis (eg, compounds being evaluated for potential therapeutic value, ie drug candidates).

  The sample or compound may also include additional components (eg, components that affect ionic strength, pH, total protein concentration, etc.). In addition, the sample may be processed to obtain at least a partial fraction or concentration. Biological samples can be stored, for example, under nitrogen, frozen, or a combination thereof, if care is taken to reduce degradation of the compound. The amount of sample used is sufficient to allow measurable detection, eg, about 0.1 ml to 1 ml of biological sample may be sufficient.

  Compounds containing candidate agents are obtained from a wide variety of sources including libraries of synthetic or natural compounds. For example, a number of techniques are available for random and directed synthesis of a wide variety of organic compounds, including biological molecules. This includes the expression of randomized oligonucleotides and oligopeptides. Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant, and animal extracts are available or can be readily produced. Furthermore, natural or synthetically produced libraries and compounds can be readily modified by conventional chemical, physical, and biochemical techniques and can be used to obtain combinatorial libraries. . Substances with known pharmacological effects may be subjected to directed or random chemical modifications such as acylation, alkylation, esterification, amidation to obtain structural analogs.

  The agent is screened for biological activity by adding the agent to at least one, usually a plurality of cell samples. This is usually done in combination with cells that do not contain the agent. Measure the change in parameters in response to the agent and compare the results with a reference culture (eg, in the presence and absence of the agent, obtained with other agents, etc.) Evaluate by comparison.

  The agent can be added to the culture medium of the cells in culture, in solution or in a readily soluble form. The agent can be added to the flow-through system continuously, intermittently, or continuously, or the compound bolus can be added to another stationary solution at once or gradually. It can also be added. In a flow-through system, two types of fluid are used: a physiologically neutral solution and the same solution with added test compound. A first fluid is passed over the cell followed by a second fluid. In the single solution method, a bolus of test compound is added to the medium surrounding the cells. The total concentration of the components of the culture medium should not change significantly by the addition of boluses or between the two solutions in the flow-through method.

  Some agent formulations do not include additional ingredients, such as preservatives, that can significantly affect the overall formulation. Accordingly, such formulations consist essentially of the biologically active compound and a physiologically acceptable carrier (eg, water, ethanol, DMSO, etc.). However, if the compound is a liquid free of solvent, the formulation may consist essentially of the compound itself.

  Multiple assays can be performed in parallel with different agent concentrations to obtain different responses to different concentrations. As is known in the art, the concentration range obtained by dilution on a 1:10 or other log scale is typically used to determine the effective concentration of an agent. The concentration can be further improved if necessary using a second dilution series. Typically, one of these concentrations, ie, zero concentration, or less than the level of detection of that agent, or the concentration of an agent that does not produce a detectable change in phenotype, or less, is a negative control. Useful as.

  Various methods are available to quantify the presence of the selected marker. Convenient methods for measuring the amount of molecules present are molecules with detectable moieties, which may be fluorescence, luminescence, radioactivity, enzyme activity, etc., especially for binding with high affinity for parameters. It is to label the molecule. The fluorescent component can be readily utilized to label virtually any biological molecule, structure, or cell type.

  The immunofluorescent component not only binds to a specific protein but can also bind to site modifications such as a specific conformation, cleavage product, phosphorylation. Individual peptides and proteins can be engineered to emit autofluorescence, for example, by expressing them as green fluorescent protein chimeras inside cells (for an overview, see Jones et.al. (1999) Trends). Biotechnol. 17 (12): 477-81). In this way, antibodies can be genetically modified to provide fluorescent dyes as part of their structure. Depending on the label chosen, parameters can be used using immunoassay techniques such as radioimmunoassay (RIA) or enzyme-linked immunosorbent assay (ELISA), homogeneous enzyme immunoassay, and related non-enzymatic techniques; Measurements can be made using other than fluorescent labels. Quantification of nucleic acids, especially messenger RNA, is also of interest as a parameter. These can be measured by hybridization techniques that depend on the sequence of the nucleic acid nucleotides. Such techniques include polymerase chain reaction as well as gene array techniques. For example, Current Protocols in Molecular Biology, edited by Ausubel et al., John Wiley & Sons, New York, NY, 2000; Freeman at al. (1999) Biotechniques 26 (l): 112-225; Kawamoto at al. (1999) See Genome Res 9 (12): 1305-12; and Chen et al. (1998) Genomics 51 (3): 313-24.

Expression Profile Database and Data Analysis A database of gene expression profiles of cancer stem cells and other cell types, and their use is also provided. Such databases typically contain various cell subpopulations (eg, cancer stem cells, cancer non-stem cells, normal counterparts to cancer cells, disease state cells (eg, inflammatory bowel cells, ulcerative) Colitis cells), virus-infected cells, early progenitor cells, early differentiated progenitor cells, late differentiated progenitor cells, and mature cells). Expression profiles and their databases can be provided in a variety of media to facilitate their use. “Medium” means a product containing expression profile information of the present invention. The database of the present invention can be recorded on a computer readable medium (eg, any medium that can be read and accessed directly by a computer). Such media include magnetic storage media (eg, floppy disks, hard disk storage media, and magnetic tape; optical storage media (eg, CD-ROM); electronic storage media (eg, RAM and ROM); and these This includes, but is not limited to, category hybrids (eg, magnetic / optical storage media) .Anyone skilled in the art will recognize that any of the currently known computer readable media records the database information of the present invention. It can be readily understood how it can be used to make a product that contains, “recorded” means a computer readable medium using any such method, as is known in the art. Means a process for storing information above, based on any convenient data storage based on the means used to access the stored information. A memory structure can be selected, and various data processing programs and formats (eg, language processing text files, database formats, etc.) can be used for storage.

  As used herein, “computer-based system” means hardware means, software means, and data storage means used to analyze the information of the present invention. The minimum hardware of the computer-based system of the present invention includes a central processing unit (CPU), input means, output means, and data storage means. One skilled in the art can readily understand which of the currently available computer-based systems are suitable for use with the present invention. The data storage means may include any product that contains a record of the information of the present invention as described above, or a memory access means that can access such a product.

  Various structural formats for input and output means can be used to input and output information in the computer-based system of the present invention. Such a presentation provides those skilled in the art with a ranking of the similarity contained in the test expression profile and identifies the degree of similarity.

  Various methods for analysis of the data set are available. In one embodiment, the expression data is converted and normalized. For example, by means of centering the expression data for each gene (by dividing the intensity measure for each gene on a given array by the average intensity of that gene across all arrays) Get. (2) The resulting ratio is then logarithmically transformed (base 2), and (3) the center is then centered on the expression data for the entire array and then the entire gene.

  For cDNA microarray data, genes with a fluorescence hybridization signal that is at least 1.5 times greater than the local background fluorescence signal in the reference channel are considered appropriately measured. Genes are centered by the average value in each data set and average linked clustering is performed.

  A graduated approach can also be employed to perform data analysis. For example, the Pearson correlation of gene expression values can provide a quantitative score that reflects the signature for each CSC. The higher the correlation value, the more similar the sample is to the phenotype of the reference CSC. Similar cells, any cell, including normal cells, progenitor cells, autoimmune phenotype cells, inflammatory phenotype cells, infected cells, differentiated cancer cells, normal stem cells, normal mature cells, etc. Can be done on the mold. Negative correlation values indicate the opposite nature. The threshold for classification can be moved above or below zero depending on the clinical purpose. For example, the sensitivity and specificity for predicting metastasis as the first recurrent event can be calculated for each threshold between -1 and +1 for the correlation score in 0.05 increments, and the desired sensitivity for predicting metastasis A threshold that results in (eg, 80%, 90%, 95%, etc.) can be selected.

  In order to obtain significance ordering, the false discovery rate (FDR) of hypotheses that are determined to be significant and that actually belong to the null hypothesis can be determined. First, create a set of null distributions of dissimilarity values. In one embodiment, the order of the observed profile values is changed to produce an out-of-chance sequence of distributions of the resulting correlation coefficients, thereby reducing the correlation coefficients. An appropriate set of distributions is made (see Tuser et al. (2001) PNAS 98,5118-21, incorporated herein by reference). A set of null distributions is obtained by: changing the order of the values of each profile for all available profiles; calculating a pairwise correlation coefficient for all profiles; correlating for this permutation Calculate the probability density function of the number; and repeat this procedure N times. Here, N is a large number, typically 300. Use the N distribution to calculate an appropriate measure (average, median, etc.) of the number of correlation coefficient values above that (similarity) value, obtained from the distribution of similar values observed experimentally at a certain significance level To do.

  FDR is the empirical data for the number of expected falsely significant correlations (estimated from correlation coefficients greater than this selected Pearson correlation in the randomized data set) Is the ratio to the number of correlations (significant correlations) greater than this selected Pearson correlation. This cutoff correlation value can also be applied to correlation between experimental profiles.

  Use the above distribution to select a confidence level for significance. This is used to determine the lowest value of the correlation coefficient that exceeds the result obtained by chance. This method is used to obtain thresholds for positive correlations, negative correlations, or both. Using this threshold, the user can filter the observed pairwise correlation coefficient values and exclude those that do not exceed the threshold. Still further, an estimate of the false positive rate can be obtained for a predetermined threshold. For each individual “random correlation” distribution, it can be found how many observations are outside its threshold range. This procedure provides a sequence of count. The average and standard deviation of this sequence provides the average number of potential false positives and their standard deviation.

  Data can be unsupervised hierarchical clustering to reveal relationships between profiles. For example, hierarchical clustering using Pearson correlation as a clustering metric can be performed. For example, correlation matrix clustering using multidimensional scaling facilitates visualization of functional homologies, similarities, and differences. Multidimensional scaling (MDS) can be applied in one, two, or three dimensions.

  This analysis can be performed in hardware or software, or a combination thereof. In one embodiment, a machine-readable storage medium is provided that is a medium containing a data storage material encoded with machine-readable data. This can display either the data set or data comparison of the present invention using a machine programmed with instructions for using the data. Such data can be used for various purposes such as drug discovery, analysis of interactions between cellular components. In some aspects, the invention is implemented in a computer program executed on a programmable computer. The programmable computer includes a processing unit, a data storage system (volatile and non-volatile memory, and / or storage elements), at least one input device and at least one output device. Program code is applied to input data to perform the functions described above, and output information is created. Output information is sent in a known manner to one or more output devices. The computer can be, for example, a personal computer, a microcomputer, or a workstation of conventional design.

  Each program may be executed in a high-level language or an object-oriented programming language to communicate with a computer system. However, the program may be executed in assembly language or machine language if desired. In any case, the language may be a compiler language or an interpreted language. Such computer programs may be stored on a storage medium or storage device (eg, ROM or magnetic disk), respectively. A storage medium or storage device may be a general purpose or special purpose for the configuration and operation of a computer when the storage medium or storage device is read by a computer to perform the procedures described herein. Read by the target programmable computer. The system may be executed as a computer-readable storage medium configured by a computer program. In this case, the storage medium so configured operates the computer in a specific and predefined manner for performing the functions described herein.

  Various structural formats for input and output means can be used to input and output information in the computer-based system of the present invention. One format for output means tests data sets with varying degrees of similarity to a reliable profile. Such a presentation provides one of ordinary skill in the art with a ranking of the similarity contained in the test pattern and identifies the degree of similarity.

Data Storage and Transmission Further provided herein are methods for computer storage and / or transmission of sequences and other data collected by the methods disclosed herein. Any computer or computer accessory, including but not limited to software and storage devices, may be utilized to implement the present invention. Sequences or other data (eg, transcriptome data) can be entered into the computer either directly or indirectly by the user. In addition, any device that can be used to sequence DNA, analyze DNA, or analyze transcriptome data can be transferred to a computer and / or storage device compatible with the computer. Can be connected to a computer. The data can be stored on a computer or suitable storage device (eg, a CD). Data can also be transmitted from one computer to another computer or to a data collection point by methods well known in the art (eg, Internet, regular mail, airmail). Thus, data collected by the methods described herein can be collected at any point or geographic location and transmitted to any other geographic location.

  An exemplary method is shown in FIG. In this example, the user places the sample into the sequencing device. Data is collected and / or analyzed by a sequencing device connected to a computer. Data can be collected and / or analyzed by software on the computer. Data can be stored, displayed (by a monitor or other similar device), and / or transmitted to another location. As shown in FIG. 10, the computer is connected to the Internet utilized to transmit data to a portable device utilized by a remote user (eg, a doctor, scientist, or analyst). It is understood that data can be stored and / or analyzed prior to transmission. In some embodiments, raw data can be collected and transmitted to a remote user who analyzes and / or stores the data. The transmission can be through the Internet, as shown in FIG. 10, but can also be through a satellite or other connection. Alternatively, the data can be stored on a computer readable medium (eg, CD, memory storage device) and the medium can be sent to the end user (eg, by mail). Remote users can be in the same or different geographic locations including, but not limited to, buildings, cities, states, countries, or continents.

Reagents and Kits Reagents and kits for performing one or more of the above methods are also provided. The reagents and kits of the present invention vary greatly. Reagents of interest include reagents specially designed for use in generating the above expression profiles of genes that are determinants of phenotype. For example, the reagents can include a set of primers for genes known to be differentially expressed in the target population or subpopulation (eg, reagents for detecting tumorigenic breast cancer cells). , Primers and probes for extending CD49f, CD24, and / or EPCAM and detecting expression).

  One type of reagent specifically tailored to generate target cell populations and subpopulation expression profiles selectively amplifies such genes for use in quantitative PCR and other quantification methods A collection of gene-specific primers designed to Gene specific primers, and methods for using them, are described in US Pat. No. 5,994,076, the disclosure of which is incorporated herein by reference. Of particular interest are at least 5 genes, often more than one of these genes, for example, at least 10, 15, 20, 30, 40, 50 Primers for 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 genes or more A collection of gene specific primers, including The gene-specific primer collection can include only primers for genes associated with the target population or subpopulation (eg, mutations, genes with known misregulations, etc.), and additional genes Primers for (eg, housekeeping genes, controls) may be included.

  The kit of the present invention may include a collection of the above gene specific primers. The kit may further include a software package for statistical analysis of one or more phenotypes, and may include a reference database for calculating the probability of sensitivity. The kit may include reagents used in various methods, such as: primers for making target nucleic acids, dNTPs and / or rNTPs (whether premixed or in separate states) One or more uniquely labeled dNTPs and / or rNTPs (eg, biotinylated or Cy3 or Cy5 tagged dNTPs), gold with different scattering spectra or Silver particles, or other post-synthesis labeling reagents (eg, chemically active derivatives of fluorescent dyes), enzymes (eg, reverse transcriptase, DNA polymerase, RNA polymerase, etc.), various buffer media (eg, hybridization buffers) And wash buffers), ready-made probe arrays, labeled probe purification reagents and components such as spin columns, signal generation and detection Reagent (e.g., streptavidin - alkaline phosphatase conjugate, chemiluminescent substance or a chemiluminescent substance) such as.

  In addition to the above components, the kit of the present invention may further include instructions for performing the method of the present invention. These instructions may be present in the kit of the present invention in various forms, and one or more instructions may be present in the kit. One form in which these instructions may exist is information printed on suitable media or substrates (eg, one or more printed information in a package insert, in a package insert). As paper). Yet another means would be a computer readable medium (eg, disk, CD, etc.) on which information is recorded. Yet another means that may exist is a website address that can be used via the Internet to access information at a remote location. Any convenient technique can be present in the kit.

  The analysis method may be embodied as a program of instructions executable by a computer to execute various aspects of the invention. Any of the above techniques can be performed by a software component loaded on a computer or other information appliance or digital device. If such is possible, the computer, instrument, or device is then used to help analyze a set of values associated with multiple genes in the manner described above, or compare such associated values. Therefore, the above technique can be implemented. The software components can be loaded from a fixed medium and can be accessed through a communication medium such as the Internet or other type of computer network. The above features are embodied in one or more computer programs that may be performed by one or more computers executing such programs.

  The following examples are provided for purposes of illustration and not for limitation.

Example 1: Analysis of gene expression in single cells A significant proportion of mouse mammary CSCs contained relatively low levels of ROS, which allowed these cells to be compared to their NTC counterparts. It was hypothesized that an enhanced level of ROS protection could be developed.

Single-cell gene expression analysis For single-cell gene expression experiments, we used a qPCR Dynamic Array microfluidic chip (Fluidigm). Single MMTV-Wnt-1 Thy1 ⁺ CD24 ⁺ Lin-CSC-enriched cells (TG) and "Not Thy1 ⁺ CD24 ⁺ " Lin "non-tumorigenic (NTG) cells were analyzed by FACS using PCR mix (CellsDirect, Invitrogen) and RNase Inhibitor (Superaseln, Invitrogen) After sorting into hypotonic lysis, we have RT-qPCR enzyme (SuperScript III RT / Platinum Taq, Invitrogen) and , A pool of low-concentration analytes (primers / probes) containing the target gene (Gclm-Mm00514996_m1, Gss-Mm00515065_m1, Foxol-Mm00490672_m1, Foxo4-Mm00840140_g1, Hifla Mm00468875_m1, Epas1-Mm00438717_m1) Reverse transcription (15 minutes at 5O 0 C, 2 minutes at 95 0 C) was followed by pre-amplification of 22 PCR cycles (each cycle: 15 seconds at 95 ° C, 4 minutes at 60 ° C). A total RNA control was also performed at the same time, each derived from one cell. Amplified cDNA was inserted into the chip sample inlet along with Taqman qPCR mix (Applied Biosystems) and individual analytes (primers / probes) were inserted into the chip assay inlet (two replicas for each). (Nanoflex, Fluidigm) loaded for 1 hour and then transferred to a reader (Biomark, Fluidigm) for thermocycling and fluorescence quantification. We excluded gene analytes whose qPCR curves showed a non-exponential increase In order to remove low quality cells (eg, dead cells) we are housekeeping genes Cells that did not express Actb (β-actin) and Hprt1 (hypoxanthine guanine phosphoribosyltransferase 1) were excluded, resulting in a total of 7 To obtain a single cell gene expression data set consisting of (non-tumorigenic tumorigenic and 139 109) cells 248 by the execution of the chip. Kolmogorov-Smirnov (KS) statistics of two samples were calculated to test whether the gene was differentially expressed in the two populations. We changed the p-value by changing the order of labeling of the samples (ie TG vs. NTG) and comparing the actual KS statistics with the statistics in the null distribution derived by changing the order. Obtained. The p-value was further corrected by Bonferroni correction to adjust for multiple hypothesis testing. Example 2 Analysis and quantification of human “colorectal cancer stem cells” (Co-CSC) using SINCE-PCR, a novel method based on “Single Cell Gene Expression Analysis”.

  The SINCE-PCR method enables the identification, characterization, and quantification of “cancer stem cells” in human colorectal cancer tissue with unprecedented purity and resolution. Cancer stem cells, which can be tumorigenic or can be tumor-initiating cells, are a subpopulation of cancer cells that can have the ability to form tumors when transplanted into immunodeficient mice. Cancer stem cell populations have now been identified in breast cancer, brain cancer, head and neck cancer, pancreatic cancer, and colon cancer. The determination and quantification of the exact function of “cancer stem cells” has several important implications for the diagnosis, prognosis, classification, and therapeutic targeting of human cancer.

  We describe a novel method for the identification, analysis and quantification of “cancer stem cells” in human colorectal cancer tissue based on single cell gene expression analysis by real-time polymerase chain reaction (real-time PCR). To do. The inventors have identified a novel set of genes that can use its coordinated and differential expression as a “signature” to identify different cancer cell subsets within the same tumor tissue. This new set of genes includes housekeeping genes common to all epithelial cells (EpCAM, β-actin, GAPDH), genes related to stem cell biology (hTERT, LGR5, survivin), and different cell lines Genes involved in related tissue-specific differentiation pathways (carbonic anhydrase II, MUC2, Trefoil family factor 3), and genes involved in the stage of normal colon epithelial differentiation (cytokeratin 20, CD66a / CEACAM1) Is included. Based on the expression pattern of this set of genes, epithelial cells purified from human colorectal cancer tissue and individually analyzed as single cells differ in their corresponding to more advanced or less advanced differentiation stages Groups (eg, terminally differentiated cells at the top of the human colon crypt versus more immature cells at the bottom of the human colon crypt), and different differentiated lineages of the colon epithelium (eg, goblet cells, enterocytes) , Immature cells) can be “classified” and clustered. Individual groups can be quantified as a percentage of the total population. We named this approach and methodology for analysis of cellular composition of biological tissues as “SINCE-PCR” (Single Cell Expression-Polymerase Chain Reaction).

  Our findings are based on several observations. Human “colorectal cancer stem cells” enriched by flow cytometry directly from freshly collected solid tumor tissue can be analyzed in a reproducible manner and reliably at the single cell level (FIG. 1).

In human colon cancer xenografts, single-cell gene expression analysis by real-time PCR has shown that EpCAM- / CD44 ⁺ cancer cells and EpCAM- / VCD166 ⁺ cancer cells are known to contain a large population of "colorectal cancer stem cells" We show that both can be further subdivided into various cell subsets characterized by coordinated and differential expression of different groups of genes involved in stem cell biology and differentiation processes. Most interesting is that a subset of cells showing higher levels for genes encoding known colon epithelial terminal differentiation markers (eg, cytokeratin 20, CD66a / CEACAM1 carbonic anhydrase II, MUC2, Trefoil family factor 3) are candidates. No or low expression of genes encoding certain intestinal stem cell markers, or genes known to be essential for stem cell function (eg, hTERT, LGR5, survivin) and vice versa Is the same. This suggests that EpCAM / CD44 ⁺ / CD166 ⁺ cancer cells contain different subsets of cells that are characteristic of various stages of differentiation (FIG. 2).

When purified by fluorescence activated cell sorting (FACS) and reinjected into immunodeficient NOD / SCID mice, CD44 ⁺ / CD66a ⁺ cells and CD44 ⁺ / CD66anewl "cells have the highest tumorigenic potential The CD44 ⁺ / CD66a ^neglow population, which exhibits properties that are similar to those of the ^{CD45 +} / CD44a ^cell population, is a differentiation marker such as CD66a / CEACAMI in the EpCAM ⁺ / CD44 ⁺ cell population. That this subset of cells characterized by high level expression of the gene encoding (ie, a subset of more “mature” cells) is often relatively depleted of tumorigenic potential. Show. On the other hand, a subset of cells characterized by no or low expression of differentiation markers such as CD66a / CEACAMI (ie, a subset of more “mature” cells) has a “colorectal cancer stem cell” content. Many.

^a それぞれの実験について、癌細胞の精製のための供給源として使用した腫瘍の異種移植片のインビボでの連続継代を、以下のように説明する:m1は、原発性の腫瘍の生着により得た1次腫瘍を示す。m2は、m1の生着により得た2次腫瘍を示す。m3は、m2の生着により得た3次腫瘍を示す。以下、段階的である。原発性は、外科検体から直接採取した原発性腫瘍を示す。
^b 選別した集団は全て、系統マーカーについて陰性である（Lin^neg）と考える。系統マーカーには、NOD/SCIDマウスにおいて確立されたヒト腫瘍異種移植片の場合においては、マウスCD45およびマウスH2-Kdが含まれ、原発性ヒト腫瘍の場合には、ヒトCD3およびCD45が含まれる（この場合、EpCAMを、陽性の上皮選択マーカーとする）。
^c 腫瘍生着は、以下のように説明する:得られた腫瘍の数/注射回数;5カ月間の経過観察後に腫瘍の塊が見られなければ、腫瘍生着は失敗とみなす。 TABLE 1 Tumorigenic properties of human colon cancer cells based on expression of CD66a / CEACAMI in combination with EpCAM and / or CD44

^{a For} each experiment, the in vivo serial passage of tumor xenografts used as a source for the purification of cancer cells is described as follows: m1 is due to primary tumor engraftment The primary tumor obtained is shown. m2 indicates a secondary tumor obtained by engraftment of m1. m3 indicates a tertiary tumor obtained by engraftment of m2. The following is step-by-step. Primary refers to a primary tumor taken directly from a surgical specimen.
^b All selected populations are considered negative for lineage markers (Lin ^neg ). Lineage markers include mouse CD45 and mouse H2-Kd in the case of human tumor xenografts established in NOD / SCID mice, and human CD3 and CD45 in the case of primary human tumors (In this case, EpCAM is the positive epithelial selection marker).
^c Tumor engraftment is described as follows: number of tumors obtained / number of injections; if no tumor mass is seen after 5 months follow-up, tumor engraftment is considered a failure.

Example 2: Generation and imaging of a human breast cancer xenograft model with lung metastases A patient-derived breast cancer specimen (chunk or TIC) was orthotopically transplanted into the breast fat pad of NOD-SCID mice. Six xenograft tumor models were created (1 ER ⁺ , 1 Her2 ⁺ , and 4 triple negative ER-PR-Her2-). All four of the triple negative xenografts produced spontaneous lung micrometastases that were evident by IHC staining (ie, staining with H & E, proliferation marker Ki67, and vimentin (Vim)). These data suggested that when transplanted into immunodeficient mice, breast tumor cells or TIC can adapt to the mouse microenvironment and repeat human tumor growth and progression with spontaneous lung metastasis.

  To facilitate dynamic semi-quantitative imaging of human breast cancer and metastasis in mice, 4 days after transplantation, mammary TIC, firefly luciferase-EGFP via pHRuKFG lentivirus (multiplicity of infection 50) The fusion gene was transduced. At the primary site, TIC could be detected by a weak biological luminescence signal. One month later, both primary tumors (with L4 and R2 breast fat pads) and lung metastases were detected and imaged with the Xenogen IVIS 200 system at the Small Animal Imaging Center of Stanford. We observed that tumor size or number of cells correlated well with signal intensity. The development of xenograft tumors with metastases and imaging of bioluminescence provide us with the feasibility of confirming MTIC miRNA function in human breast cancer in vivo in this proposal.

Example 3: Human breast MTIC microarray and real-time PCR analysis human breast primary cancer source cells (TIC) or metastatic ^{TIC (MTIC) (CD44 + CD24} - / low ESA + lineage -) a, breast primary Isolated from site or pleural effusion. Once lung metastases were detected in the xenograft model, the lungs were separated with blenzyme (Roche) and the MTIC population (CD44 ⁺ CD24 ^{− / low} ESA ⁺ H2K ⁻ , FIG. 8a) was purified. Cells were stained with mouse H2K and human CD44, CD24, and ESA. The MTIC population grows orthotopic tumors at a rate of 5/8 after transplantation into mouse mammary fat pads when using 200-1000 sorted cells.

  Target genes regulated by HIF1α and HIF1, including Snail, Zeb2, Vimentin, E-cadherin, Lox, Cox2, and VEGF, compared to non-tumorigenic tumor cells, as shown by microarray analysis and real-time PCR It was differentially expressed in MTIC (FIG. 8B). Co-localization of HIF1α, vimentin, and CD44 was confirmed by immunohistochemical staining.

Example 4: MicroRNA analysis MicroRNA screening identified differential expression profiles of original breast cancer stem cells and metastatic cancer cells isolated from the lung. For example, miR-10a expression was higher and levels of miR-490, miR-199a, etc. were lower in lung MTIC than in primary breast TIC. Indicated by triple real-time PCR in Figure 4, comparison of the average CT value of the lungs MTIC for primary TIC: miR-1Oa (-7.9) , miR-490 (+ 3.0), and miR-199a ⁽⁺ 12.9). NR3 was used as an internal control. The data showed that miR-1Oa was up-regulated 27'9-fold and miR-199a was down-regulated 212-9-fold in MTIC over breast cancer primary TIC.

Example 5: CD66a as a non-tumorigenic cancer cell marker for breast cancer
Breast cancer cells were sorted based on CD44 and CD66a, while the majority of cells were CD244bw. Cells were then transplanted onto the breast fat pad of NOD / SCID mice and tumor growth was monitored. CD44 / CD66a-cells showed higher engraftment rates as well as faster growth rates due to bioluminescence imaging, as shown in FIG. CD66 ⁺ cells show lower tumor growth rates and slow tumor growth rates, tumor sizes are much smaller, and very similar flow profiles compared to CD66-derived tumors It was.

In FIG. 5a, the flow profile is shown based on the CD44 and CD66a markers. CD66-CD44 ⁺ cells and CD66 ⁺ CD44 ⁺ cells were sorted for in vivo tumorigenicity assay (100 cells or 1000 cells were placed on Zid or 41h of NOD / SLID mouse mammary fat pad) Transplanted). As 10b shows, 5 out of 8 transplants derived from 100 CD66− cells grew into tumors, while those derived from 100 CD66 ⁺ cells grew only 2 out of 8. For 1000 cells, 8 out of 8 injected with CD66- cells grew, but only 3 out of 8 with CD66 ⁺ cells grew into tumors. Comparing the growth rate of palpable tumors, CD66 ⁺ cells were much slower and smaller in size than those generated from CD66 − cells (FIG. 5c). In FIG. 5d, 100K CD66-CD44 ⁺ cells or CD66 ⁺ CD44 ⁺ cells were infected with firefly luciferase-EGFP lentivirus prior to injection. The biological luminescence signal by CD66 ⁺ cells was higher than that of CD66 − cells at the start (day 13). However, after one or two months, CD66-cells showed an overt bioluminescent signal and eventually grew into a palpable tumor (day 68).

Example 6: Optimization of the gene list used to identify and measure the frequency of cancer stem cells Most markers used at this point to identify both normal and cancer stem cells are essential stem cell functions Is not related. Their expression is related to the specific microenvironment in which the stem cells are present at the time of isolation. Thus, the usefulness of common markers used to identify stem cells can vary depending on the site from which they are collected.

  Our approach was to identify important stem cell function markers. Since the typical property of stem cells is self-renewal, we focused our efforts on the regeneration pathway. We have identified a number of genes that are highly expressed by normal HSCs, leukemia stem cells generated from progenitor cells, and human epithelial cancer stem cells, but not by cells that do not self-renew in their respective tissues. Was identified. This genomic analysis described in previous results has identified several genes that have been previously associated with stem cell self-renewal. Similarly, the inventors have identified candidate microRNAs that are differentially expressed by breast cancer stem cells and non-tumorigenic cancer cells. Evidence indicates that some of these genes and microRNAs have important stem cell functions, and that the functions of these genes are also important for hESC and iPSC self-renewal and maintenance .

  In order to obtain a device that can measure the frequency of cancer stem cells in a tumor cell population, it is desirable to optimize the gene list used to identify cancer stem cells. As shown in FIG. 1B, the inventors have made great progress on this and have identified telomerase as a cancer stem cell marker, as well as several genes associated with the process of self-renewal. The telomerase component TERT is expressed only in colon cancer cells having an immature phenotype. Furthermore, TERT is not efficiently downregulated with the differentiation of some hESC and iPSC lines.

  Both normal and cancerous colonic epithelial cells are analyzed for expression of genes associated with crypt cell maturation and self-renewal. The list of self-renewing genes is expanded beyond TERT to maximize the confidence that the cell is a stem cell. The gene expression of normal stem cells and cancer stem cells identified in the analysis by the present inventors is measured. Because cancer stem cells can possibly arise from either stem cells that have escaped restrictions on growth or progenitor cells that have escaped the counting mechanism that limits the number of mitosis they can undergo, the candidate gene is normal It is a gene expressed by mouse HSC, mouse leukemia stem cells derived from progenitor cells, and human breast cancer stem cells. The first candidate genes identified in this list (which are all related to stem cell maintenance) include BMI1, -IDI, IGFBP3, HOX family members H0XA3, H0XA5, MEIS1, ETS1, ETS2, RUNX2, And STAT3. We confirm which of these genes are associated with cancer stem cell self-renewal. To do this, we systematically test these candidate genes for their role in cancer stem cell self-renewal using in vitro and in vivo techniques.

  Expression of genes that regulate self-renewal is related to the expression of genes specific to epithelial cells, including maturation markers such as keratin and intestinal mucin. Thereby, it is considered that it is possible to confirm that the cell under analysis is not a normal cell mixed in the biopsy material. Mutations in the tumor suppressor gene whose expression is down-regulated by the self-renewal gene BMI I allow early progenitor cells to self-renew. These genes are frequently mutated in colon cancer, so that self-regenerating colon cancer stem cells are thought to arise from both normal stem cells and early colon progenitor cells. Still further, oncogenic mutations alter gene expression by colon cancer cells. Thus, there may be differences in the expression of at least some genes associated with early crypt cell maturation between normal colonic epithelial stem cells and their malignant counterparts. Thereby, these two self-regenerating cell populations can be distinguished from each other.

  The inventors have identified 37 miRNAs that are differentially expressed in cancer stem cells and non-tumorigenic cancer cells. Some miRNA clusters were down-regulated in normal tissue stem cells but not down-regulated in cancer stem cells. Furthermore, the expression of some miRNAs (eg miR-200c and miR-183) suppresses the growth of embryonic cancer cells in vitro, eliminates their tumorigenic potential in vivo, and in vitro of breast cancer cells The ability to form clones was inhibited. These miRNAs, and other clusters that we have identified, provide the molecular link that connects the biology of breast cancer stem cells and normal stem cells. Expression of these microRNAs, which were consistently up-regulated or down-regulated in tumorigenic cells, derived from undifferentiated and differentiated hESCs and iPSCs Probe in a single cell. In essence, undifferentiated cells are sorted by cell surface markers distinct from pluripotent stem cells (eg, Tra and SSEA subtypes), miRNA expression, replate efficiency, and in vivo population parameters (embryonic phase). Cancer, mixed embryonal carcinoma / differentiated cell index (% EC vs. differentiated), and teratoma assay results for differentiated cells) are evaluated. Differentiated stem cell populations are obtained by production of embryoid bodies and sorted by positive and negative selection for SSENTRA markers after 28 days of differentiation. The inventor tests single cells in the selected population for: (1) a microRNA profile indicative of cancer stem cells, (2) a gene expression profile (below), and (3) transplant / teratoma assay. result. We have found that "resistant to differentiation" cells in these populations form malignant embryonic cancer derivatives, differentiated cells in single cells and undifferentiated cells. It was expected that cellular markers would be expressed simultaneously.

Example 7: Gene expression profile at the single cell level In a population of pluripotent cells, we still down-regulate important markers of tumorigenicity (eg TERT) even after 21 days of differentiation. Observe strains that cannot rate (see Figure 6). In addition, we are unable to down-regulate both exogenous and endogenous pluripotency markers in the differentiated state, approximately 50% of our iPSC lines. Observed that. In essence, this suggests a “molecular war” between differentiation and self-renewal that allows us to predict the outcome of a tumorigenic tendency. We optimize the gene list for the identification of malignant cells in hESC and iPSC cell cultures by: (1) undifferentiated hESC and IPSC, and human Assay genes overexpressed in EC (embryonic cancer) cells compared to embryonic blastomeres, (2) include genes from Aim 1 (for cancer stem cell identification) Cross-reference the list of genes, and (3) adding differentiated somatic and germline genes (the latter remains in pluripotent stem cells that are resistant to differentiation). We then use an immunodeficient mouse assay to assess the tumorigenic potential of a subpopulation diagnosed according to the likelihood of being malignant based on the underlying single cell gene expression.

  CNV analysis. Chromosomal variants, including frequently observed loss and gain of chromosomes, are associated with instability in the pluripotent human stem cell population. However, few studies are working on more dense, high-throughput methods to assess copy number at multiple loci. We propose to tailor our technique for the assessment of genome-wide CNV numbers in separately derived pluripotent stem cell lines. Changes in CNV are thought to reflect subchromosomal instability. First, we have a special probe to add to our gene / locus list that recognizes duplication for the entire genome, including what we have observed so far in our laboratory. The set can be designed (Figure 6). SCAD can be adapted to analyze up to 1000 markers in its initial design. CNV analytes are commercially available and may correlate with genomic instability in hESCsIiPSCs.

Example 8: Operation of an automated device to identify and quantify cancer stem cells Based on a combination of cell surface phenotype and gene expression, the cancer stem cells are automatically identified and their frequency in the tumor calculated. Design the device. Using the optimized marker / gene analysis described herein, a similar strategy could be malignant based on the co-expression of differentiated and undifferentiated markers in single cells. Used to identify a cell. This device creates single cell suspensions of embryoid bodies or tumor needle biopsies, isolates subpopulations of cells (epithelial cells, differentiated cells, undifferentiated cells), and then hundreds or QRT-PCR is performed on a thousand single cells to determine the stem cell content of a tumor or pluripotent cell culture. Such a fully automated device eliminates the laborious steps currently required for flow cytometric sorting of cancer stem cells and isolates sufficient cancer cells for PCR assays to quantify cancer stem cells. In addition, it would be possible to enable a clinical diagnostic tool that requires less than 100,000 cells and requires no manual intervention. The automated operation, effectiveness, and low cost associated with microfluidic chip technology are expected to enable rapid genetic diagnosis tailored to the individual.

  At the heart of this system is the microfluidic cell sorter. This device isolates viable cells (epithelial cells or cultured pluripotent cells, or their products) from debris (necrotic cells and other particles) and uses fluorescent signals from up to five different surface markers. Cells are then sorted from single cell suspensions and placed in individual bottles for subsequent genetic studies. Other upstream steps such as digesting the tumor or cell culture to obtain a cell suspension and staining the cells with fluorescent surface markers can be incorporated into the system. A method for using this system for tumor analysis is described herein. Once the biopsy material is obtained, the physician places the sample in the input port of the system. Using an easy-to-handle computer interface, the physician sets the parameters required for sorting and genetic analysis, such as the number and type of surface markers and the number of PCR cycles required. The machine then performs the remaining steps without human intervention. Based on the techniques described above, the system would be capable of sorting at least 30 cells / second.

  A single cell analysis device (SCAD) can be modular (Figure 7) and performs the following steps in an integrated, fully automated manner: (1) Tissue digestion: Place the tissue in the input port of the device. The appropriate enzyme is loaded into the device and circulated to digest the extracellular matrix to obtain a cell suspension. (2) Separation of viable cells from debris: The suspension typically includes viable cells having an average size of 10 to 15 micrometers and debris material having an average size of approximately 5 micrometers. Is included. The amount of dead material is often relatively large compared to viable cells, so it is important to filter out dead material for efficient isolation of cells . We do this by flowing the digested tissue suspension through a microfluidic “metamaterial”. Microfluidic metamaterials can divide fluid flow according to particle size. (3) Staining: The filtered single cell suspension is stained using appropriate surface markers in different compartments of the microfluidic device. Staining with up to five different markers can be useful in obtaining a highly pure population of cancer cells. (4) Sorting: The stained single cell suspension is circulated to the next compartment of the microfluidic device to sort out cancer cells from the remaining cells. Poisson statistics and Monte Carlo simulations show that only 2,000 to 20,000 cancer cells are needed to sort to detect a 2-fold change in cancer stem cells with 99% confidence. Such a small number of cells currently cannot be sorted efficiently using flow cytometry. This is because the initial sample size required for FACS is approximately 1 million cells. We do this using microfluidic-based sorting that circulates the cell suspension indefinitely in an airtight isolated small volume environment without wasting cells.

  Flow-based microfluidic cell sorter: A microfluidic cell sorter with a throughput of approximately 50 cells / second has been demonstrated, where cells are circulated at high speed by laser light (Di Carlo et al. Lab Chip 2006; 6 : 1445-1449) to detect and analyze scattered light. High speed electronics and more efficient imaging equipment can improve processing power by an order of magnitude, reducing sorting time to less than 10 minutes.

  Parallel sorting: A cell sorter is developed based on high density capture of cells on a 2D array of microfluidic chambers that can be individually addressed (FIGS. 7B and 7C above). Cells are circulated through a sorter array and captured by a microfabricated basket. Such baskets have previously been shown to have a single cell capture efficiency of more than 50% in a free flowing suspension (Di carlo et al., Supra). After filling all baskets, the microfluidic valve is closed and the array is designed in a specially designed, computer-controlled light in all five fluorescent colors needed to identify tumorigenic cells. Make an image. The new chip is also capable of phase contrast imaging, which can prove useful for studying cell morphology. The identified tumorigenic cells are circulated to the next module for lysis, while the rest of the cells are circulated out of the chip. This new cell sorter can work with very few initial cell numbers. This is because the cells can be circulated many times and are therefore considered not to be wasted. Current microfluidic chip technology can place nearly 10,000 of these elements in a 3 × 3 cm area. These can be quickly queried (in a single image) using state-of-the-art imaging devices such as those used by the Fluidigm Biomark system. This cell sorter has a throughput of approximately 30 cells / second. One advantage of using a parallel sorting device as opposed to a flow-based cell sorter is that the imaging between sorting and PCR can be performed by the same imaging device, which Can relate data on fluorescence and morphology to genetic data of individual cells.

  Cell lysis and mRNA capture: Sorted cancer stem cells are circulated to the next module for lysis in individual chambers. mRNA was captured on a row of oligo-dT beads, reverse transcribed on the beads as previously shown (Marcus et al. Anal Chem 2006; 78: 3084-3089), and a new developed for Heliscope Reagents for processing chips by gene sequencing protocols or transferring to macroscopic wells (in microliter range) and pre-amplifying a set of genes according to the protocol of the invention below Can be mixed. Pre-amplified samples are transferred to modules similar to Fluidigm Dynamic array chips for qRT-PCR and determination of true cancer stem cell content.

Based on the analysis of normal breast and blood stem cells, as well as colon cancer stem cells, head and neck cancer stem cells, and breast cancer stem cells, we have identified a novel single cell analyte. This makes it possible for the first time to accurately and unambiguously identify and count cancer stem cells in biopsy materials and cultured pluripotent stem cell populations. As a proof of principle, we applied this assay to the analysis of single colon cancer cells. To this end, we used FACS to screen CD66 ⁺ CD44 lineage colon cancer cells from early passage xenografts established from two different patients. These markers can be enriched approximately 3 to 5 times the colon cancer stem cells (CoCSC) in the tumor. We suspected that the cancer cells isolated using these markers were only partially enriched for CoCSC. Single cell gene expression analysis and subsequent tumorigenicity studies show that the actual CD66 ⁺ CD44 ⁺ lineage cells are a mixture of CoCSC and non-tumorigenic cells, and using this assay, biopsy specimens It shows that the frequency of CoCSC in can be identified more accurately. Single cell analysis reveals a stratified hierarchical developmental structure of colon cancer cells resembling normal colon crypts. Of note, we have found that most immature cells in colon tumors express TERT, one component of the telomerase complex that is important for long-term tumor maintenance. The expression of LGR5, which is indicative of normal colon stem cells, is also limited to immature cells. In contrast, genes expressed by mature colon crypt cells (including MUC2, 'CK20, CA-2, and in particular CD66a) are markers for immature cells (most notably TERT). It was expressed by cells that were not co-expressed. This suggests that these cells are limited in their ability to undergo active mitosis as well as epithelial crypt cells in normal maturation. In fact, we transplanted CD66a ⁺ (differentiated colon cancer cells) and CD66a ′ ″ colon cancer cells into immunodeficient mice. CD66a 'cells formed tumors (5 of 6 injections), but CD66 ⁺ cells did not form tumors (0 of 5 injections). Similarly, in two human breast cancer tumors tested, CD66ew cells were enriched for cancer stem cells when tested in an immunodeficient mouse model. These results indicate that single-cell gene expression analysis can identify and quantify cancer stem cells in biopsy materials and cultures.

Example 9: Gene expression signature shared by normal and cancer stem cells in both blood and breast epithelial tissue Recently, it has become clear that cancer stem cells can arise from different cell compartments. Some probably arise from mutant stem cells that are off-limits on the expansion of the pool of stem cells. Others arise from more differentiated early progenitor cells that have lost the counting mechanism that normally limits the number of mitosis they can undergo. Of course, many markers of cancer stem cells or leukemia stem cells arising from stem cells or progenitor cells are different. However, regardless of the cell from which it originates, stem cells retain the ability to self-renew. We share some of the pathways that regulate self-renewal in cancer stem cells arising from either the stem cell compartment or partially differentiated progeny, and also share normal HSCs. It was judged that there was a possibility. To test this hypothesis, we expressed by normal mouse progenitor cells (ie, self-regenerating populations) that are expressed by mouse leukemia stem cells (ie, self-renewing populations) derived from normal mouse HSCs and progenitors. It was analyzed whether genes that are not expressed by non-regenerating populations) are also expressed by human breast cancer stem cells, but not by their non-tumorigenic counterparts. Of note, human cancer stem cells overexpress these genes, but not their non-tumorigenic counterparts (FIG. 8). We have also created two other gene lists to identify other potential candidates: (i) expressed by breast cancer stem cells and normal breast stem cells, but non-tumorigenic Genes not expressed by cancer cells or mature breast epithelial progenitor cells, (ii) expressed by normal human HSCs and human breast cancer stem cells, but not by human blood progenitor cells or non-tumorigenic breast cells gene.

  Many of these genes are associated with self-renewal and cancer. These include insulin growth factor binding partner IGFBP3, HOX family members H0XA3, H0XA5, ME1S1, and transcription factors such as ETS1, ETS2, RUNX2, and STAT3. We tested whether the transcription factor STAT3 is a true cancer stem cell regulator. STAT3 plays a role in the maintenance of both ES cells and HSCs. Genomic analysis of both mouse and human breast cancer stem cells revealed that many STAT3-activated transcripts were overexpressed by cancer stem cells. Next, when we tested immunochemical analysis of breast tumors, cells that were STAT3 positive tended to be enriched at the invasive edge of the cancer, and this protein was found in the internal part of the tumor. It was not seen in cells that looked more differentiated. Finally, there are small molecule inhibitors of STAT3. Such inhibitors can be tested in a cancer stem cell model. The effect of the STAT3 inhibitor cucurbitacin on the colony forming ability of mouse breast cancer stem cells was tested. Short-term exposure to this inhibitor for 24 hours reduced the number of colonies by approximately 50% (p <0.02, t-test). These results suggest that STAT3 plays an important role in at least some breast cancer stem cells.

  The second gene of interest is MEIS1. MEIS1 is selectively expressed by normal blood and breast stem cells, leukemia stem cells, and breast cancer stem cells. Genetic studies indicate that MEIS1 expression is absolutely necessary for the self-renewal and maintenance of both normal blood stem cells and their leukemia counterparts. MEIS1 can regulate the regeneration of breast cancer stem cells.

  Particularly interesting candidate genes expressed by both normal and cancer stem cells include: CAV1, GAS1, MAP4K4 (kinase), MYLK (kinase), PTK2 (kinase), DAPK1 (kinase), LATS ( Kinase), FOSL2, AKT3 (kinase), PTPRC (tyrosine phosphatase), MAFF (oncogene), RRAS2 (related to RAS), NFKB, ROBO1, IL6ST (activates STAT3), CR1M1, PLS3, SOX2, CXCL14 , ETS1, ETS2, MEIS1, and STAT3, and CD47. Interesting candidate genes that are overexpressed by cancer stem cells but not overexpressed by normal stem cells include RGS4, CAV2, MAF (oncogene), WT1 (oncogene), SNAI2, FGFR2, MEIS2, 101, 103, ID4 , And FOXC1 are included.

Example 10: Whole Transcriptome Analysis of Hematopoietic Stem Cells In this example, we seek to use hematopoietic stem cell transcriptome analysis. A general overview of this embodiment is shown in FIG. In this example, a population of cells suspected of containing hematopoietic stem cells is isolated from a test subject. The cells are then prepared for FACS analysis by exposing the cell population to a fluorescent antibody against a known hematopoietic stem cell marker (eg, CD34, Thy1, etc.). Cells are sorted into 96-well plates so that each well does not contain more than one single cell.

  The isolated single cells are lysed and the lysate is divided into two parts. For the first part, distinguish HSC and non-HSC by either expression level or presence of expression (eg CD34 +, CD19-, CD17-) essentially as described in Example 1. Single-cell gene expression analysis is performed by real-time PCR using a selection of genes that can. After identifying HSCs within the population, lysates from single cells identified as HSCs are pooled. A cDNA library is generated by amplifying total mRNA using standard methods. The cDNA is then sequenced using any “next generation” method as described herein. The sequenced transcriptome is then analyzed to determine whether a unique gene and / or surface marker is present.

  After identification of surface markers specific for HSC, antibodies that specifically bind to the surface markers are prepared by commercially available techniques. Confirm the specificity and effectiveness of the antibody (eg, binding to isolated and / or recombinant protein). The antibody is then labeled with a fluorescent component. Subsequently, FACS sorting and / or analysis can be performed on other populations of cells (eg, from the same subject or different subjects) using antibodies to the newly discovered surface antigen.

Example 11: Analysis of therapeutic agents In this example, selection of candidate therapeutic agents is performed. Target cells (eg, colon cancer stem cells and colon cancer cells (differentiated) are isolated and analyzed at the single cell level as described above. Target markers specific to the previously identified target cells (eg, target cell specific FACS separation using fluorescent antibodies and / or fluorescent labeling of target cell specific nucleic acids) to separate target cells from biopsy specimens.

  Target cells are separated into addressable locations containing single cells. The isolated cells are then exposed to a library of candidate therapeutics (eg, antibodies, toxin-binding antibodies, small molecules). Cells are then collected and analyzed for gene expression patterns and / or cell viability. An effective candidate therapeutic can be one that targets cells to kill. Alternatively, candidate therapeutics are those of genes known to be misregulated (eg, upregulated or downregulated) compared to expression patterns from cells that are not disease states. Expression can be altered. Exposure of the target cell to the candidate therapeutic can result in a change in the expression pattern of the nucleic acid that more closely resembles that of a normal (ie, not in disease state) cell. Candidate therapeutics that are promising in killing or altering target cells are then exposed to normal cells to determine their potential for use as therapeutics (e.g., candidate agents are targeted cells and If normal cells are killed, they can be excluded from potentially useful agents).

Claims (45)

A method of identifying different cell populations in a heterogeneous solid tumor sample comprising the following steps:
Randomly distributing individual cells from the tumor to discrete locations;
Performing a transcriptome analysis on a plurality of genes of cells individually distributed to the distinct locations; and performing a cluster analysis to identify one or more different cell populations.   2. The method of claim 1, wherein the individual cells are not enriched prior to the dispensing step.   2. The method of claim 1, wherein transcriptome analysis is performed on at least 1000 individual cells simultaneously.   2. The method of claim 1, wherein the transcriptome analysis is performed using nucleic acid analysis.   2. The method of claim 1, wherein the discrete locations are on a planar substrate.   2. The method of claim 1, wherein the randomly dispensing step is performed in a microfluidic system.   2. The method of claim 1, wherein transcriptome analysis comprises analyzing expressed RNA, non-expressed RNA, or both.   2. The method of claim 1, wherein the transcriptome analysis is a full transcriptome analysis.   The method of claim 1, wherein transcriptome analysis comprises amplifying RNA using a set of primer pairs.   10. The method of claim 9, wherein the primer pair is not a nested primer.   The method of claim 1, wherein transcriptome analysis is performed on all or a subset of individual cells simultaneously or substantially in real time.   2. The method of claim 1, wherein the one or more cell populations are normal stem cells, normal progenitor cells, normal mature cells, inflammatory cells, cancer cells, cancer stem cells, or non-tumorigenic stem cells. A method for analyzing a heterogeneous tumor biopsy material from a subject comprising the following steps:
Randomly distributing cells from the biopsy material to separate locations;
Performing a transcriptome analysis on at least 50 genes of the individually distributed cells; and identifying one or more characteristics of the tumor using the transcriptome data.   14. The method of claim 13, wherein the performing step is performed without prior enrichment of the cell type.   14. The method of claim 13, wherein the characteristic is the presence or absence of cancer cells or the number.   14. The method of claim 13, wherein the characteristic is the presence or number of stem cells, early progenitor cells, early differentiated progenitor cells, late differentiated progenitor cells, or mature cells.   14. The method of claim 13, wherein the property is the effectiveness of a therapeutic agent in eliminating one or more of the cells.   14. The method of claim 13, further comprising using the characteristic to diagnose a subject having cancer or a stage of cancer.   14. The method of claim 13, wherein the property is signal transduction pathway activity.   20. The method of claim 19, wherein the signaling pathway is specific for cancer stem cells, differentiated cancer cells, mature cancer cells, or combinations thereof. A method for identifying a signaling pathway utilized by a diseased cell comprising the following steps:
Randomly distributing cells from a heterogeneous sample;
Performing a transcriptome analysis on the distributed cells;
Identifying at least one disease state cell using transcriptome analysis;
The transcriptome analysis of cells of the at least one disease state;
(A) cells that are not in a disease state;
(B) a cell in a different disease state; and (c) comparing to a transcriptome of a stem cell in a disease state; and
Signal transduction that is expressed in (i) cells in disease states, (ii) stem cells in disease states, and (iii) depending on the situation, in cells in different disease states, but not in cells that are not disease states Identifying the pathway, thereby identifying the signaling pathway utilized by the diseased cell.   24. The method of claim 21, wherein the disease state is cancer, ulcerative colitis, or inflammatory bowel disease.   24. The method of claim 21, wherein the signaling pathway is required for survival of the diseased cell. A method for diagnosing a symptomatic subject comprising the following steps:
Randomly distributing cells from a heterogeneous sample;
Performing a first transcriptome analysis on the distributed cells;
Using transcriptome analysis, the first transcriptome analysis of cells in at least one disease state is compared to a second transcriptome analysis of cells that are not in the disease state Identifying cells in a state, thereby diagnosing the presence or absence of symptoms associated with the cells in the disease state in the subject.   25. The method of claim 24, wherein the disease state is breast cancer, colon cancer, ulcerative colitis, or inflammatory bowel disease.   23. The method of claim 21, wherein the transcriptome analysis comprises analyzing expressed RNA, non-expressed RNA, or both.   24. The method of claim 21, wherein the transcriptome analysis is a full transcriptome analysis.   24. The method of claim 21, wherein the diseased cell is a breast cancer stem cell. A method for screening for therapeutic agents comprising the following steps:
Exposing a first subject having cells in a disease state to one or more test agents;
Obtaining a heterogeneous tumor biopsy material derived from the subject from a region of interest;
Performing a transcriptome analysis on at least one individual cell from the heterogeneous tumor biopsy, wherein the biopsy comprises cells in one or more disease states; and ,
The transcriptome analysis is
(I) a second subject having no diseased cells; or (ii) comparing to a transcriptome from any of the first subjects prior to the exposing step; and
Identifying an agent that affects the transcriptome of the cell from a test region that is more similar to the transcriptome of the second subject or the first subject prior to exposure. A method of determining the potential effect of a therapeutic agent on a disease comprising the following steps:
Dividing a first population of cells in a disease state into individual locations, the individual locations comprising individual cells;
Determining the expression level of at least one nucleic acid or protein from at least one of said individual cells, thereby creating an expression signature of the disease state;
Exposing a second population of diseased cells to the agent;
Dividing the second population of cells in the disease state into individual locations, wherein the individual locations comprise individual cells;
Determining the expression level of at least one nucleic acid or protein from at least one individual cell from the second population; and the expression level from the individual cell from the second population; Comparing to the expression signature of the disease state, thereby determining the effect of the agent on the disease.   32. The method of claim 30, wherein the exposing step is performed in vivo.   32. The method of claim 30, wherein the first population and the second population are isolated from the subject.   40. The method of claim 32, wherein the subject is a human.   32. The method of claim 30, wherein the disease is cancer, ulcerative colitis, or inflammatory bowel disease.   32. The method of claim 30, wherein the nucleic acid or protein is a cancer stem cell marker.   32. The method of claim 30, wherein the expression level is an mRNA expression level.   34. The method of claim 33, wherein determining the mRNA expression level comprises detecting expression of 10 or more nucleic acids or lack of expression.   32. The method of claim 30, wherein the expression level is a protein expression level.   32. The method of claim 30, wherein the dividing step comprises exposing the population of cells to an antibody that specifically binds to a protein present on the individual cells. A method for determining the likelihood of a subject's response to a therapeutic agent, including the following steps:
Dividing a population of cells from a subject into individual locations, the individual locations comprising individual cells, wherein at least one of the individual cells is a diseased cell;
Determining the expression level of at least one nucleic acid or protein from at least one of the individual cells of the disease state, wherein the nucleic acid or protein is the target of a therapeutic agent; and
Determining the likelihood of a reaction by the subject based on the expression level of the at least one nucleic acid or protein.   41. The method of claim 40, wherein the expression level is an mRNA expression level.   42. The method of claim 41, wherein determining the mRNA expression level comprises detecting expression of 10 or more nucleic acids or lack of expression.   43. The method of claim 42, wherein the expression level is a protein expression level.   43. The method of claim 42, wherein the dividing step comprises exposing the population of cells to an antibody that specifically binds to a protein present on the individual cells.   41. The method of claim 40, wherein the therapeutic agent is an anticancer agent.

Images